The premise sounds like a gimmick and turns out to be an engine. You are Simon P. Jones, fourteen years old, and your great-uncle has left you a house called Mount Holly on one condition: find Room 46. The estate has forty-five rooms. The floor plan is a grid, five wide and nine deep, with the entrance hall at the bottom and a sealed antechamber at the top. Forty-five rooms, forty-six needed. That’s the whole hook, delivered in the first sixty seconds, and it takes most players a very long time to work out what kind of question it actually is.

Blue Prince arrived in April 2025 from Dogubomb — essentially Tonda Ros — published by Raw Fury, on PC, PlayStation 5 and Xbox Series consoles. It has been described as a puzzle game, a roguelike, a deduction game and a walking sim, and the reason nobody can settle on a label is that its central mechanic belongs to a genre that doesn’t have a name yet. You don’t explore Mount Holly. Youdraft it.

Drafting as level design

Here’s the loop. Open a door, and rather than a room, you get three blueprints: pick one, and it becomes the room behind that door, permanently for this run. Each blueprint shows its footprint — how many doors it has, and which walls they’re on — plus its cost and any rules attached. Some rooms only place in dead ends. Some cost gems. Some can only appear in the outer columns, or in the back half of the estate. Some are rank-limited, appearing once and never again.

Then you walk in, and the room does something. It might contain a key. It might contain a lever, a note, a shop, a security terminal, a slot for a coin. It might contain nothing except three more doors and three more decisions.

The resource that governs all of it issteps. You start each day with a step budget, every room you enter spends one, and when the steps run out the day ends, the house empties, and tomorrow’s Mount Holly is a fresh sheet of paper. Everything you built is gone.

Set that beside how roguelikes normally work and the difference is sharp. InDead Cells orHades, the run generates a level and you react to it. InBlue Prince the run generatesoptions and you author the level from them, which makes every door a small architectural argument with yourself. Do I take the room with four exits because I need the reach, or the room with one exit because it has a chest in it and I’m nearly out of steps? Do I place the Boiler Room now, knowing it’s cheap and useless, because placing anything else here costs gems I don’t have?

The genius is that the drafting rules are themselves the puzzle. A room that only spawns in dead ends means dead ends have value. A room that must sit on the west wall means the west wall is a resource. Within a few hours you stop seeing a floor plan and start seeing a constraint satisfaction problem with wallpaper. I can think of no other game where the act ofbuilding the dungeon is the intellectual content and the act of walking through it is the reward.

The step economy is the difficulty curve

The elegance of steps as a currency is that it prices everything at once. Curiosity costs steps. Backtracking costs steps. A room that turns out to be a dead end with a locked door costs a step going in and, if you routed badly, several more getting back out. There is no health bar and nothing kills you; the only enemy is the walk itself.

That makesBlue Prince one of the very few games wherelayout efficiency is the skill. Good players don’t have faster reflexes. They have a better sense of the grid — they know that placing a corridor at row three buys them lateral movement for the rest of the day, that a room with doors on three sides at the bottom of the map is worth more than the same room at the top, that spending eight steps on a detour to a shop is only correct if you already have the coins. The difficulty curve is invisible because it’s inside your own planning, and it flattens the moment you get better at reading the grid, which is the most honest kind of progression there is.

The genuine cruelty is that Mount Holly is stingy with the thing you need most, which is reach. Room 46 sits at the top of the grid. Getting there requires an unbroken chain of drafted rooms from the entrance hall to the antechamber, which requires door alignment, which requires luck, which requires that you spend your entire day building a corridor rather than looting one. The house is constantly offering you interesting rooms that lead nowhere and boring rooms that lead north, and choosing correctly means choosing boredom over and over. That’s a real design risk, and the game takes it deliberately.

What actually persists

If the house resets every day, what carries? Two things, and the split is the reason the game works.

The first is a modest layer of permanent unlocks — keys, codes, tools, changes to what can appear in the drafting pool. It’s real, it’s slow, and it’s the least interesting part.

The second isyou. What actually persists across days is the notebook in your head. The house is stuffed with documents: letters, ledgers, memos, a newspaper, plaques, timetables, a set of family records that don’t agree with each other. Read them and a second game emerges underneath the drafting one, made of numbers you can’t use yet, names that mean nothing yet, and rules that turn out to be literal. The most powerful thing you can take out of a run is a fact.

This is the same trickReturn of the Obra Dinn andLorelei and the Laser Eyes run: the gate is knowledge, and knowledge doesn’t reset.Blue Prince welds that to a roguelike’s structural churn, so the randomness that would ruin a fixed puzzle game becomes the delivery mechanism for the clues. A run that ends four rooms short of the antechamber still hands you three documents, and the documents are the actual progress. Once you understand that, the failed days stop feeling like failures. You aren’t trying to reach Room 46 today. You’re trying to learn something today, and Room 46 falls out of enough somethings.

I’d argue the real ancestor is the old cassette-era adventure: the games I typed into a C64 in the eighties where you kept the map on graph paper because the machine wasn’t going to keep it for you, and the memory the game relied on was yours.Blue Prince is that idea rebuilt with a modern designer’s understanding of variance. The graph paper is back. Get a real notebook — the in-game journal does some of the filing, and it doesn’t do the thinking.

Where it fights itself

Two things chafe, and both come from the same source.

Variance can hand you a genuinely dead day. Not a hard day — anothing day, where the drafts come up cheap and doorless, you’re out of gems by row three, and you walk the corridor already knowing there’s no route north. Ten minutes of a game that has no combat and no failure state is ten minutes of walking. The design’s answer is that documents still drop, and the answer holds most of the time. It doesn’t hold all of the time.

And the late game asks a lot. Once you’ve cracked the surface,Blue Prince keeps going — considerably further than most players expect — into puzzles that assume you’ve been transcribing details for thirty hours and cross-referencing them off-screen. That’s not a flaw so much as a filter, and the game is admirably unbothered about who it filters out. It won’t tell you when you’re done. It won’t tell you that you missed something. It just leaves it there.

The verdict

Blue Prince is the rarest thing in games: a mechanic nobody has done before, executed by someone who understood exactly what it was for. The drafting isn’t a delivery system for the puzzle box; the draftingis the puzzle box, and the manor is the physical form of a decision tree. It is also gentle, funny, beautifully lit, and quietly sad about the family it’s describing — a game that could have been a pure abstraction and chose to be a house instead.

The step economy will frustrate anyone who wants a puzzle game to sit still and be solved. Everyone else gets forty hours of the specific, disreputable joy of realising that a note you skimmed on day six was an instruction. Play it on PC if you want the notebook open on a second screen; the console versions play identically and you’ll just want paper instead.

If it lands, go toLorelei and the Laser Eyes next for the same respect for your attention, orAnimal Well for the same conviction that the real progression happens in your head.

Spoilers below

The moment the game shows its hand is when you realise the forty-five-room count was never the constraint. Mount Holly’s grid has edges, and the game spends its opening hours training you to treat those edges as walls. They are not. Once the estate proves it can extend past its own footprint, the drafting rules you’d internalised as physics turn out to be conventions, and every “impossible” placement you’d written off becomes a question again.

That’s the structural rhyme with the story. The house is a document about a family that lied about its own shape — an inheritance built on a boundary that was drawn wrong on purpose. Simon is handed a floor plan and a puzzle, and the puzzle is that the floor plan is a claim, not a fact. Learning to distrust the grid and learning to distrust the paperwork are the same act, arriving at the same time, which is about as tight as a game and its theme ever get.

The most aggressive thingPersona 5 Royal does to you happens on an ordinary Tuesday in June. School finishes. The game returns control. You have one afternoon and one evening, and in front of you sit a part-time job that would raise your Charm, a friend who has been waiting three in-game weeks to advance a relationship, a book that would raise your Knowledge before an exam, and a dungeon with a deadline in eleven days. You can do one afternoon thing. Some of those options will eat the evening as well.

Nothing threatens you. No enemy is on screen. This is the most tense the game ever gets, and Atlus knew that when they built it.

The system, stated plainly

Persona 5 came out in Japan in September 2016 and in the West in April 2017, directed by Katsura Hashino, with Shigenori Soejima’s character design and Shoji Meguro’s soundtrack doing an enormous amount of the heavy lifting.Royal is the expanded edition — Japan in October 2019, the West in March 2020 on PS4, and eventually everywhere in October 2022 when Atlus finally put it on PC, Xbox, Switch and PS5.

The structure is a school year. Each day gives you a small number of discretionary slots. You spend them on Confidants — twenty-odd relationships, each ranked one to ten, each granting mechanical benefits as it climbs — or on the five social stats, Knowledge, Guts, Proficiency, Kindness and Charm, which gate the Confidants. Meanwhile the plot delivers Palaces: cognitive dungeons with a hard calendar deadline. Fail to finish one by its date and the game ends. Actually ends.

So every single day is an allocation problem with an audit at the end of it. That is the machine. The phantom-thief business, the jazz, the extraordinary menus — all of it is upholstery on a spreadsheet, and the spreadsheet is where the feeling comes from.

Why the scarcity works

Scarcity is easy to design and hard to makehurt. Most games with a time limit produce anxiety, which is a cheap emotion.Persona 5 produces regret, which is expensive.

The mechanism is that the game makes the thing you gave up visible. You skipped Yusuke, who texted, and whose Confidant is at rank four, and whose rank five would give you a combat ability you can name and want. The cost of every choice is a person with a face and a text message. That is the whole trick, and it is why the strategy-game version of this system, where you allocate abstract workers to abstract buildings, does not feel like anything.

The second mechanism is that the game refuses to let you win the allocation. A completionist run of Royal is documented to sit well past a hundred hours, and even a perfect one is a scramble, because the Confidant ranks are gated behind social stats that are themselves gated behind days. You are always behind. The design’s honest position is that a year is too short to be good at everything, which is a considerably more mature statement than the phantom-thief plot manages in its entire runtime.

The real ancestor of this is not the RPG lineage at all. It isTokimeki Memorial, Konami’s 1994 dating simulator, which established the loop of stat-raising against a school calendar with affection as the scoring function.Persona 3 imported that structure into Atlus’s demon-fusion RPG in 2006 and discovered that the calendar made the dungeonmatter, because now the dungeon had an opportunity cost. EverythingPersona 5 does is the third refinement of that fusion. Fifteen years earlier, on the home computers I grew up with, the closest equivalent wasElite’s fuel economy — a resource that turned exploration into a decision — and the shared idea is old: the interesting number is the one you cannot have more of.

Where Royal fights itself

Here is the argument I want to make about the definitive edition, and it is not a comfortable one.

Royal’s improvements are almost allloosening. Palaces now contain Will Seeds, which reward exploration with SP-restoring accessories; SP was the original’s scarcest combat resource and the reason a Palace took multiple visits. The grappling hook opens shortcuts. Safe rooms let you leave and return without losing progress. Morgana’s early curfew — the notorious business of a cat sending you to bed — is relaxed. Ranked SP items are purchasable. Showtime attacks give you a free burst of damage on a random timer.

Each of these is a quality-of-life win. Collectively they mean that a Palace which used to consume three or four calendar days now goes down in one, and those recovered days go straight back into your Confidant budget. Atlus made the dungeons kinder and thereby made the calendar softer, and the calendar was the antagonist.

I do not think this ruinsRoyal. I think it is a real cost that reviews at the time under-reported because the additions were so obviously generous. The 2016 game’s cruelty was doing structural work: when SP ran out, the Palace ended, and you went home having spent a day and achieved nothing, andthat was the day you learned what the game was about.Royal rarely gives you that day. It replaces it with a third semester — new content from November onwards, a new Confidant in Takuto Maruki, a new party member in Kasumi Yoshizawa — that is, and I say this having sat with it for years, the best-written material Atlus has ever shipped and structurally the wrong place to put it.

Because the third semester arrives after the calendar has finished threatening you. It is a coda. It is superb, and it is playing in a mode the entire preceding hundred hours had been arguing against: a stretch of time where the pressure is narrative rather than arithmetical. The best thing inRoyal is the part ofRoyal that stops beingPersona 5.

The upholstery, briefly

I have been dismissive about the surface and should correct that, because the surface is why anyone tolerates the spreadsheet.

The interface is the most confident work in the medium. Every menu is animated, angled, red-and-black, scored, and no two transitions in the game are the same shape. That is functional work: a game asking you to spend a hundred hours in menus has to make the menus a place you enjoy standing, and Atlus solved that by treating the pause screen as a piece of graphic design rather than a list. Meguro’s soundtrack does the same job in the other channel: the battle theme is the reason a random encounter you have fought two hundred times still snaps you awake.

And the writing, when it is good, is very good indeed. Futaba’s Palace, in particular, does something the genre almost never manages: it takes a character’s psychology, renders it as architecture, and then makes navigating the architecture the act of understanding her.

When it is bad it is very bad. The plot’s politics are adolescent, the Phantom Thieves’ celebrity arc goes nowhere it has not been dragged, and there are stretches of the middle third where the game will hold you in a cutscene for forty minutes to say something it said at the start.

The verdict

Persona 5 Royal is a hundred-hour game about the arithmetic of a finite year, wearing the best clothes in the industry. Its greatness is entirely structural — the fact that a Tuesday in June is harder than any boss — and its expanded edition is a slightly compromised version of that greatness in exchange for content nobody could reasonably decline.

PlayRoyal, since it is the only version anyone can now buy, and play it on whatever is nearest; the 2022 ports run fine and the PC one is the obvious pick. Take the third semester. Do not look up an optimal Confidant route, because a schedule someone else calculated removes the entire game.

For a much smaller, much sharper treatment of the same idea — time as a currency you are always short of, with none of the upholstery — readCitizen Sleeper, which does in eight hours what this does in a hundred and is honest about which of those is a virtue. And for the other end of the JRPG’s structural ambition,Nier: Automata makes the shape of the playthrough itself the argument.

Spoilers below

Maruki is the reasonRoyal exists, and he is a better antagonist than Shido by a distance that is almost embarrassing.

The third semester’s premise is that the counsellor who has been helping you all year has acquired the power to remove suffering, and has used it, and the world you wake up in is one where everybody’s worst thing simply did not happen. The dead parent is alive. The ruined career is intact. The friend’s trauma is gone, and so is the person the trauma made.

What makes this land is that it is acalendar argument. Maruki’s reality is a place with no opportunity cost — every choice is available, nothing is foreclosed, no Tuesday in June ever charges you anything. He is offering the player exactly what the player has spent a hundred hours resenting the absence of. The game has trained you to want this. The whole system has been a machine for generating the specific hunger Maruki proposes to satisfy.

And the refusal — the choice to reject a painless world and go back to the one where things cost — is the only moment in the entirePersona series where the theme and the mechanics say the same sentence at the same time. Everything else inPersona 5, the rebellion, the masks, the chains, is a metaphor stapled to a combat system. This is not a metaphor. This is the game asking whether you understood what the schedule was for.

The answer arrives in February, in a fistfight in the sky, which is the most Atlus thing imaginable and does not diminish it in the slightest.

Final Fantasy VII arrived in Europe in November 1997, and I remember the specific texture of that moment better than I remember most of the game: the Amiga had gone quiet under me, the PlayStation had turned up carrying the future on three discs, and a Japanese role-playing game about a mercenary with a comically large sword was suddenly the thing everyone had an opinion about. Midgar — the opening city, the bombing run, the plate, the slums — took maybe six hours. It was the best six hours on the console and then the game left it behind and went off to be a world map.

Final Fantasy VII Remake, released for PS4 in April 2020, takes those six hours and makes them the whole game. Forty of them, give or take. This is the fact that decided most people’s reaction before they had played a minute of it, and it is the least interesting fact about the thing.

The combat is the achievement

Start here, because the combat system is the part that deserves study, and it is a genuinely clever piece of hybrid engineering rather than a compromise.

You control one party member directly in real time. Attacking builds ATB — the gauge from 1997, now a resource rather than a clock — and spending ATB is how you cast, use an item, or fire an ability. Opening the command menu drops the game into Tactical Mode, where time slows to a crawl while you choose. You can switch to any party member instantly, and the ones you are not controlling fight competently and build ATB slowly on their own.

The read: this is an action game whose action generates the currency of a turn-based game, and whose turn-based game happens inside a bubble of slowed time carved out of the action. It sounds like a fudge. In play it produces something quite precise — you are always simultaneously executing and planning, and the slow-motion menu functions as a held breath.

Then Stagger. Every enemy has a stagger gauge, filled by pressure — pressure being generated by doing the specific thing that enemy dislikes, which the scan spell will tell you. Staggered, they take a burst of extra damage for a window. This mechanic did not come fromFinal Fantasy VII. It came fromFinal Fantasy XIII, and the people who built it are the people who made this: Motomu Toriyama on scenario, Naoki Hamaguchi on game design, both veterans of the XIII trilogy, under Tetsuya Nomura’s direction.XIII was pilloried for a combat system that most players never got far enough in to understand.Remake takes that system’s actual idea — combat as the management of a vulnerability window rather than an exchange of numbers — and wraps it in something you can touch.

That is the real ancestral trace, and it is more honest than the one the marketing wanted.Remake’s fighting isKingdom Hearts hands welded toFinal Fantasy XIII’s brain, made by the people who had both. The 1997 game contributes the Materia system, which returns almost unchanged and remains one of the best build systems anyone has designed: spells are objects, objects go in slots, slots are in gear, and therefore your entire capability is portable between characters at any moment.

Where the design fights itself

Forty hours of Midgar is not, in itself, a crime. Midgar can hold forty hours; the concept is strong enough and the art direction is comprehensively magnificent. The problem is thetexture of the added thirty-four, and specifically that Square Enix appears to have solved a large fraction of them with corridors.

There is a category of moment in this game that everyone who played it can describe: the gap you squeeze sideways through while the level streams, the ladder you climb at a fixed speed, the arm you hold out to a robot claw, the pair of levers that must be pulled in sequence in a room built exclusively to contain two levers. The Train Graveyard chapter and the long descent through the Shinra building are the usual examples cited, and they are cited because they are correct. These are pure duration, purchased at the cost of your goodwill.

I take this seriously because forty hours is a real thing to ask. A game asking for a working week of somebody’s life owes them a reason for every one of those hours, andRemake has stretches where the reason is plainly that the chapter needed to be longer than the content in it. The side quests are the same instinct wearing a friendlier face — go and find some cats, go and kill some rats — and they exist to buy you affection scenes and a slightly better relationship with characters the game could have simply written more scenes for.

What rescues it, and it does mostly rescue it, is that the expansion isspecific where it counts. Jessie, Biggs and Wedge are three names and a death in the original. Here they are people with a flat, a family, opinions about Cloud, and a plan; when the plate falls the arithmetic has changed, because you have had dinner with them. Wall Market is no longer a joke and a dress — it is an economy, with Don Corneo sitting on top of it and a genuinely superb set piece built around a piece of 1997 comedy that could have gone very badly and instead goes big, sincere, and slightly magnificent.

The thing it is actually doing

Here is the read that matters. Every remake has to decide what it is faithfulto, and the available options are the text or the memory of the text.Dead Space’s 2023 rebuild chose the text and treated the job as restoration — clean the varnish, fix the joints, do not repaint.Capcom’s Resident Evil 4 chose to argue with its original about tone, and won some of that argument.

Final Fantasy VII Remake chose something stranger and considerably riskier. It is faithful to thememory — to the way Midgar is bigger in your head than it ever was on the disc, to the way Aerith is more important in retrospect than she is in her introduction — and it treats that gap between the game and the recollection as the actual subject. The forty hours are the game rendering your inflated memory at the size your memory has it. That is why the added material is nearly alltexture rather than plot: the plot was never what got exaggerated.

And then it goes further, in a way I will keep below the line.

The verdict

This is a magnificent, sincere, overlong thing, and the overlength follows directly from the ambition. Play it. Play it on PC or PS5 where Intergrade’s performance mode makes the combat legible at the speed it wants to run — the PS5 version arrived in June 2021 with the Yuffie episode attached, and the PC release followed later that year, reaching Steam in 2022. Turn the difficulty to Normal and use Assess on everything; the combat only opens up once you accept that scanning is a verb.

And then decide, honestly, whether you have another eighty hours forRebirth, because the 2024 sequel doubles everything here including the problems.

If you want the argument about endings and structure taken somewhere weirder,Nier: Automata is Square’s other great swing at making the shape of a playthrough into a statement.

Spoilers below

The Whispers. Let us have this out.

From very early on,Remake has ghosts in it — spectral figures who appear whenever the story deviates from the events of 1997 and physically shove it back on course. They stop Aerith from saying a thing. They block a road. They intervene at the moment Barret dies in a way he did not die before, and undo it.

The obvious reading is that they are the game’s own canon, personified and armed. The obvious complaint is that this is Nomura being Nomura: a metafictional device bolted onto a story that did not need one, resolved in a final act where the party fights the concept of fate in a sky arena while a dead character from a spin-off walks past.

I have gone back and forth and I have landed here: the Whispers are the honest expression of what this project is. Square Enix could not remakeFinal Fantasy VII straight — the original exists, is playable on everything, and is better than a straight remake would be at the only thing a straight remake could offer. What they could do is make a game about the pressure of the original’s existence, in which the characters are pushed around by a force that wants events to go the way you remember. The Whispers are the audience. They are the wiki. They are the decades of accumulated players who know Aerith dies and will riot if she does not.

And the ending — the party choosing to fight that force, and the game closing on a Midgar that has explicitly diverged from the one on the discs — is a studio saying out loud that it refuses to be a museum. I think the execution is muddled. The sky arena is bad. The Sephiroth escalation arrives about six hours before it has been earned. But theidea is the boldest thing a major publisher has done with its own back catalogue, and the alternative — a respectful, tasteful, identicalFinal Fantasy VII with better shaders — would have been a much safer game and a much emptier one.

The thing you remember is not the thing that happened.Remake knows that, and builds the knowledge into the plot. Whether it can land the consequences across three games is a question for the third one.

Somewhere around the second act ofPathfinder: Wrath of the Righteous, the game stops asking you what kind of hero you are and starts asking what kind ofthing you are willing to become. It offers, among other options, becoming a lich. This goes considerably past an evil alignment tick-box: the game grows a new questline, a new set of abilities, a new attitude from your companions, and a substantially different ending. It does this for ten separate answers to the question.

Owlcat Games released this in September 2021, three years afterKingmaker, funded by a Kickstarter that pulled in around two million dollars. It is adapted from Paizo’s tabletop adventure path of the same name and built on Pathfinder’s first-edition ruleset, which is to say the ruleset that took Dungeons & Dragons 3.5 and asked what would happen if nobody ever said no. That question is the entire aesthetic of this game, and I want to defend it.

The sheet is the toy

Twenty-five base classes. Archetypes on top of those in numbers that require a wiki. Multiclassing with no meaningful guard rails. Feats that combine into things the designers plainly did not sit down and enumerate. If you have spent any time in thePathfinder build community you will know the folk canon: the Scaled Fist monk dip that turns your charisma into armour, the Vivisectionist alchemist who is a rogue with better chemistry, the Sword Saint whose entire job is to make one attack per round mathematically obscene.

The honest criticism of this is that it is not balanced, and the honest answer is that balance was never the promise.Wrath of the Righteous is not a game about a curated encounter budget. It is a game about a character sheet as a construction kit, where the pleasure is the pleasure of ademoscene release — someone found a hole in the rules and drove a lorry through it, and everybody else gathered round to admire the lorry.

The real ancestor here is SSI’s Gold Box run on the eight- and sixteen-bit machines.Pool of Radiance in 1988 gave you a party, a rulebook, and no interest whatsoever in whether your combination of decisions was sensible; the manual assumed you would read it and the game assumed you had. Everything that came after — theBaldur’s Gate line, and especiallyThrone of Bhaal’s High Level Abilities, where 2000-era BioWare quietly conceded that the endgame of a build game is absurdity — descends from that permission. Owlcat picked the permission back up after twenty years of the genre apologising for it.

What the mythic paths actually do

The headline is the ten mythic paths, and it is worth being precise about their mechanical shape, because “branching narrative” undersells it.

You commit around Act 2. Aeon polices causality and can retroactively unmake things. Azata is a chaotic good party that summons a dragon called Aivu who grows up over the campaign. Demon eats your enemies’ powers and your own restraint. Trickster rewrites the game’s jokes into rules — it can make a critical hit on a Perception check literally see through the plot. Lich raises your dead enemies and quietly poisons every companion relationship you have built. There is also Gold Dragon, Angel, Devil, Swarm-That-Walks and Legend, which discards mythic power entirely to become spectacularly good at ordinary things.

The design read: a mythic path is alens, applied at the midpoint, that recolours content you were going to see anyway. That is an enormously efficient piece of engineering. Owlcat did not build ten campaigns. They built one campaign with ten sets of rules about how you are allowed to interact with it, and because the paths land at Act 2 rather than at character creation, you have already met everyone before the lens goes on. Your companions therefore have opinions about the change. Regill, the hellknight, approves of order arriving from any direction. Arueshalae, the succubus trying to stop being one, has a complicated time watching you become a demon.

This is the thingWrath does that no other CRPG of its era matches, including the one that sold ten times as many copies.Baldur’s Gate 3 has better faces, better cameras, better everything you can photograph.Wrath of the Righteous has more consequential branching per pound, and the branching is welded to the build rather than sitting beside it.

The crusade, and the argument about it

Then there is the army layer. In between the dungeons,Wrath hands you a strategic map, a recruitment budget, stacks of crusader units, and a set of turn-based battles that play like a thriftyHeroes of Might and Magic. It is optional in the sense that you can set it to resolve automatically. It is not optional in the sense that it sits in the middle of the game asking for your time.

Most reviews at the time called it a mistake. I think that is too quick, and the reason is thematic rather than mechanical. The adventure path is about acrusade — a mass mobilisation of ordinary people against a demonic incursion — and every CRPG convention pushes against that theme, because the CRPG convention is that six exceptional individuals resolve everything personally. The crusade layer is the game’s way of insisting that the war exists whether or not your party is in the room. When your mythic power scales up the units you can field, the two layers finally touch, and the campaign map stops being homework.

The problem is the tuning. The economy is opaque, the general system rewards a couple of obviously correct picks, and the layer arrives before you have any reason to care about it. Owlcat’s own answer — the Enhanced Edition update in 2022, and the option to skip it — is a studio admitting the theory outran the execution. I would rather a studio reach for something structurally interesting and land it two-thirds of the way than ship the safe version.

Where it fights itself

The launch was rough. This is a matter of public record and Owlcat spent the following year patching it hard: scripting faults, encounter bugs, save issues, the usual toll of a game with this many interacting systems shipped by a mid-sized studio. It is in a substantially better state now, with the Enhanced Edition and five DLC releases behind it, and the console versions arrived in 2022 with the fixes baked in.

The deeper structural fault is the difficulty.Wrath is a game whose default settings assume you have read the rulebook, and whose enemies from Act 3 onwards start layering resistances, spell immunities and attack routines that will simply erase an unoptimised party. The customisable difficulty sliders are the most important feature in the game and the least advertised: you can tune enemy stats, damage taken, and the swinginess of the maths independently, and doing so is the correct response rather than an admission of anything. The record puts a full run somewhere well past a hundred hours, and a hundred hours is a real thing to ask of someone’s life. Any game asking it should let the asker set the terms.

And it is long past the point of shame in the fourth act. The Midnight Isles content, the drift into demon-realm sameness, the sheer volume of trash encounters between the good ideas — this is a game that would be better at eighty per cent of its length and does not believe that for a second.

The verdict

Wrath of the Righteous is the most generous CRPG of the last decade, and generosity is a real virtue with real costs. It gives you more rules than you can hold, more branches than you can see in one run, and more army admin than you asked for, and it does so out of a conviction that the player is an adult who can decline the parts they dislike. That conviction is rarer than good writing and considerably rarer than good faces.

Play it on PC where the mods and the build community live. Turn the difficulty sliders to something honest before Act 3 rather than after it. Pick a mythic path that scares you slightly, because the game is at its best when the answer to “what are you becoming” is one you are not entirely comfortable with.

For the opposite approach — an RPG where the character sheet has almost no combat function at all and is instead a set of arguments you have with yourself — seeDisco Elysium. The two games are as far apart as the genre stretches, and they are both right.

Spoilers below

The Trickster path is the one I want to single out, because it is a joke that turns into a thesis.

Trickster’s mythic abilities are structured as punchlines: improving your Perception until you notice things the plot did not intend you to notice, improving Knowledge until enemies are humiliated by trivia, turning the rules themselves into a bit. It reads as the comedy option for about ten hours. Then the game starts quietly showing you what a person who treats a demonic invasion as material actually looks like from the outside, and the companions start noticing, and the ending has a cost the jokes were papering over the entire time.

The Lich path does the inverse and does it more cruelly. It is the strongest path mechanically and the loneliest narratively — the game gives you power and then removes, one by one, the people who liked you. Several companions leave. One in particular can be kept only by doing something to them, and the game does not soften what that is.

The Aeon path’s best moment is the retroactive one: the ability to declare that a thing which happened did not, applied to a specific historical injustice the campaign has already presented as settled fact. It is the single most CRPG thing in the game — a rules interaction that is also a moral position — and it exists because Owlcat took Paizo’s mythic rules literally instead of politely.

That is the case for maximalism. A restrained version of this game would have had one ending, four classes and no lorry-sized holes in its rules, and nobody would still be arguing about it five years later.

Before you accept a mission inDeep Rock Galactic, the game tells you how long it will take. This has nothing to do with the vague estimated-campaign-length figure on a store page. On the mission select terminal there is a row of little icons, one to three, for length, and another row for complexity, and they mean what they say. A one-dot mission is a quarter of an hour. A three-dot is closer to half. You can look at the wall, look at the clock, and make an informed decision about whether you have time for this before the washing machine finishes.

I have been trying for years to explain why that small piece of interface design makes me trust Ghost Ship Games more than any studio operating a seasonal calendar, and the honest answer is that everything else about the game follows from it.

The four verbs

Ghost Ship, a Copenhagen studio, putDeep Rock Galactic into early access in February 2018 and shipped 1.0 in May 2020. The premise is that you are one of four space dwarves employed by an aggressively cheerful mining corporation to go into procedurally generated caves on an alien planet, extract something, and get out before the bugs finish eating you.

The design decision that makes it work is that each of the four classes owns a way of moving through the world, rather than owning a role in a combat triangle. The Driller carries a pair of drills that eat tunnels through solid rock, and he is the reason a vertical shaft becomes a ramp. The Engineer plants platforms out of a gun and strings ziplines, and he is the reason an unreachable mineral vein is now a floor. The Gunner fires ziplines across chasms and drops a bubble shield that pauses the world for six seconds. The Scout has a grappling hook and a flare gun, and he is the reason anyone can see anything at all.

Every cave is fully destructible. Put those two facts next to each other and the whole thing clicks: the map is a problem, and each of you holds a different tool for deforming the problem. When a mission goes well it is because four people independently reshaped the same rock into something navigable without ever discussing it. When it goes badly it is because the Driller has tunnelled somewhere private, the Scout is two hundred metres up a wall, and the Gunner and the Engineer are having a nice quiet time being eaten.

The real ancestor of this is not the co-op shooter lineage at all. It isBoulder Dash — Peter Liepa and Chris Gray’s 1984 C64 game, one of the first things I ever loaded off tape that made me think of terrain as a material rather than a backdrop.Boulder Dash understood that digging is a verb with consequences, that the tunnel you cut is a tunnel that exists afterwards, and that the tension in a mining game comes from the geometry you yourself created.Lemmings took the same idea onto the Amiga seven years later and made it about other people’s stupidity.Deep Rock Galactic is the version where the terrain you ruined is a shared social space, and the stupidity is yours.

Why the loop holds

The mission structure is where the craft is. A Mining Expedition asks for a quota of Morkite; Point Extraction wants Aquarq crystals hauled to a central pad; Salvage has you defending stationary uplinks on a timer; Escort Duty walks a drilling machine called Doretta through the cave and dares you to keep her alive. They are all, structurally, the same shape — go in, do a task under pressure, call the drop pod, run for it — and yet they do not blur, because the task changes what the cavemeans.

In a Mining Expedition the cave is a larder and you wander it greedily. In Salvage the cave is a defensive perimeter you have five minutes to understand. In Escort Duty the cave is a corridor being carved in front of you by something that does not care about your opinion. Same rocks, same bugs, same four dwarves. Completely different reading of the space. That is the trick that lets a game with a handful of mission types stay legible across hundreds of hours, and it is a far more efficient use of design effort than shipping thirty modes.

Then there is the extraction. The drop pod lands, a timer starts, and now every mineral in your pack is a bet against your ability to sprint. This is the single best-tuned moment in the game and it is essentially free: it costs Ghost Ship nothing to add a countdown, and it converts the last ninety seconds of every mission into a small farce. Somebody always dies. Somebody always has to be carried. The dwarf who mined the most is invariably the one who is furthest away when the door opens.

Nitra is the other quiet masterpiece. It is the mineral that buys resupply pods, at eighty per pod, and it is scattered like everything else. Which means your ammunition economy is amining problem. Run dry and the answer is to go find some rock. No shop, no loadout screen mid-mission, no crate. The resource that keeps you shooting is the resource you are already there to dig, and the two systems fuse instead of sitting next to each other.

The part everyone else should copy

Deep Rock Galactic runs seasons. Season 01 arrived in late 2021 and they have kept coming. Here is what a season is: free, for everyone, with a progression track you can work through by playing normally. Here is what happens when it ends: the cosmetics and gear from that season roll into the general loot pool, permanently available to anyone who shows up afterwards.

Read that again, because the industry standard is the exact inverse. Nothing expires. Nothing is held hostage behind a date you missed. There is no paid tier of the pass. The studio sells optional cosmetic packs, and that is the whole monetisation. A player who buys the game in 2026 can obtain everything a 2021 player has, by playing.

This is a coherent theory of what a co-op game is for, and charity has nothing to do with it. Ghost Ship appears to have concluded that the product is the fifteen minutes in the cave, and that anything which makes those fifteen minutes feel like an obligation is damaging the product. The Deep Dives, the weekly three-stage runs on a shared seed, are the one concession to a calendar, and even those are a treat rather than a tax: a special hard thing that expires, sitting on top of a permanent library that does not.

The social furniture matters here too, and it is easy to be sniffy about. The salutes. The “Rock and Stone!” shout mapped to a button. The Abyss Bar on the space rig where you can drink beers that apply modifiers and dance badly at a jukebox. Every one of those is a small mechanism for turning four strangers into a crew before the mission starts, and they work on a population of anonymous players in a way that no amount of voice-chat etiquette guidance ever has. It is the most consistently pleasant public lobby in the genre, and that is a design achievement; the demographics did not manage it on their own.

Where it fights itself

The gear progression is slow in a way the rest of the game is not. Overclocks — the build-defining weapon modifiers — come from Deep Dives and machine events and a forge with a randomised cost, and the randomisation means the specific overclock you want may simply decline to appear for weeks. A game this generous with its content is oddly stingy about its builds.

And Hazard 5 is where the design’s honesty runs out slightly. The lower hazards are a physics comedy; Hazard 5 is a game about knowing the spawn logic, and the gap between them is a small cliff wearing the costume of a ramp. The custom difficulty settings added later paper over this, and they are the right answer, but they arrived years after the players who bounced off had already gone.

The verdict is thatDeep Rock Galactic is the best-structured co-op game of its generation, and the structure is the argument. It is on PC, PlayStation and Xbox, has been on Game Pass, and asks for a fifteen-minute commitment and nothing else. Play it with three people you like. Play it with three strangers; that works too, which is the point.

If you want to see the same studio’s ideas run through a different mill,Deep Rock Galactic: Survivor takes this world and pours it into the auto-shooter shape I wrote about inthe Vampire Survivors review, and it is a much better licensed spin-off than it had any need to be. For the other end of co-op — the one that punishes dawdling instead of scheduling around your evening —Risk of Rain 2 is the counter-argument.

Spoilers below

A game with no plot has little to spoil, though it has a shape worth knowing about.

The season storylines — the Rival Corporation’s robots, the caretaker sequence, the various things Mission Control declines to explain — are told almost entirely through voice lines, terminal barks and the occasional new enemy that turns up without introduction. This is Ghost Ship’s most underrated decision. The narrative is ambient, skippable, and never once stops a mission to deliver itself. You learn that a rival company has been mining the same rock because you found their machinery in your cave, got shot by it, and had to deal with that instead of watching it.

Compare that to what a live-service game usually does with story, which is to interrupt your fifteen minutes with a cutscene about a war you did not enlist in. Ghost Ship’s version respects the same rule as the mission timer: your evening is yours, the cave is the game, and anything the studio wants to tell you has to fit around the digging.

And then there is Karl. Nobody will tell you who Karl was. The dwarves toast him, blame him, invoke him. The joke has run for years and the studio has never explained it, which is exactly the right call — Karl is a folk practice, and folk practice dies the moment somebody writes the lore page.

There is a version of this review that is four words long and reads “it costs four quid”. Poncle’sVampire Survivors has been on Steam since December 2021, went 1.0 in October 2022, and has spent the years since being the cheapest thing on anyone’s account that they have still somehow put a hundred hours into. It won the BAFTA for Best Game in 2023, which remains the funniest sentence in recent awards history — a browser game built in Phaser by one bloke in Italy, sat in the same category as budgets with a comma in them.

The joke has been told. What has not been examined nearly enough is the actual engineering. BecauseVampire Survivors is one of the cleanest pieces of subtractive design I have seen since the sixteen-bit years, and the thing it subtracted is the thing every other game in its lineage assumed was load-bearing.

The removal

You move. That is the input. The left stick, or WASD, and nothing else. Your weapons fire on their own timers, in their own patterns, forever, whether you are paying attention or not. Monsters come in from the edges in tides. Every kill drops an experience gem; every few gems is a level; every level throws up a small menu of weapons and passive items and asks you to pick one. Thirty minutes on the clock, and at thirty minutes Death arrives and takes the run off you regardless of what you have built.

Strip a twin-stick shooter of the aiming and you would expect to be left with nothing. Instead you are left withpositioning, and it turns out positioning was carrying the whole genre the entire time. The direct ancestor here is well documented — Galante has been open that he took the shape fromMagic Survival, a Korean mobile game from 2021 — but the deeper root is the arcade lineage that runs throughRobotron: 2084, where the second stick was really just a way of expressing which pile of enemies you had decided to be nearest to.Vampire Survivors noticed that the interesting decision in that genre was always the standing — the shooting was bookkeeping — and had the nerve to delete everything else.

What this does to the moment-to-moment is peculiar and specific. You stop reading the screen as targets and start reading it as terrain. A whip that fires horizontally means you want enemies on your flanks; a King Bible orbiting your body means you want to be inside the crowd rather than backing away from it; a Garlic aura means the correct play is to walkinto the thing that is killing you. Your build silently rewrites what “safe” means, and half the skill of a run is noticing that the geometry of safety changed two levels ago and you are still moving like the old build.

The economy underneath

The level-up menu is where the design does its real work, and it is worth being precise about why it lands, because “you get a choice of upgrades” describes a hundred games that feel like nothing.

You have six weapon slots and six passive slots. Weapons cap at level eight. A capped weapon plus the correct passive item unlocks anevolution, which arrives only from a chest dropped by a boss after the ten-minute mark. That is three separate resources — slots, levels, time — all converging on a single delivery window, and the effect is a run that has a genuine dramatic structure rather than a difficulty slope. Minutes zero to ten are you deciding what the run is about. Ten to twenty are the evolutions cashing in and the screen turning to soup. Twenty to thirty are you finding out whether the soup is thick enough to survive what is coming.

The screen-turning-to-soup is the part people describe as the game playing itself, and it is where the criticism usually stops. I would argue the opposite: the soup is thereward, and it is a reward the design has to earn by making the first ten minutes genuinely precarious. EarlyVampire Survivors, before your first evolution, is a horrible tense scrabble in which one bat can end you and the gems are always eight steps too far away. The power fantasy at minute twenty-two is only legible because you remember being nearly killed by a bat at minute two. Take away the fragile opening and the whole thing collapses into a screensaver, which is precisely what happens in the dozens of imitators that let you start strong.

Galante’s professional background before this — years designing for the gambling industry — is a matter of public record and he has talked about it openly, and you can see the fingerprints without needing to be rude about it. The gem-collection radius, the near-miss, the drip of small rewards between the big ones: this is somebody who knows exactly which frequency the reward loop wants to run at. The difference is that a slot machine’s schedule exists to extract, and this one exists to keep a decision interesting. Same tooling, aimed somewhere better.

Where it fights itself

The meta-progression is the weak joint. Gold from runs buys permanent PowerUps — more might, more speed, more luck, more revives — and it does the standard roguelite thing of making early runs artificially bad so that later runs can feel earned. It works, in the sense that it kept me coming back. It also means that a substantial chunk of your first several hours is spent losing to a wall that exists because a spreadsheet says you have not paid yet.Hades, which I have written abouthere, got away with this by attaching a story beat to every failed run;Vampire Survivors attaches a number.

The Arcana system, added at 1.0, is the more interesting late addition and the more uneven one. Arcana cards are run-wide rule changes — one makes your first weapon fire on a timer independent of cooldown, another converts recovery into damage — and the good ones are genuinely build-defining in the wayBalatro’s jokers are, a comparison I have leaned on before inthat review. The problem is that a few of them are so much better than the rest that the “choice” of two Arcana at the start of a run is frequently no choice at all. When your deck contains a card that trivialises the ceiling, the ceiling stops being a place you visit.

And the content has sprawled. Poncle has kept adding — Legacy of the Moonspell in late 2022, Tides of the Foscari in 2023, anAmong Us crossover, aContra one, and eventually the official Konami-licensedOde to Castlevania in 2024, which retroactively legitimised an aesthetic the game had been cheerfully gesturing at since the itch.io days. Each pack is generous and cheap. Collectively they have turned a game whose original virtue was that you could understand all of it into one with a collection screen you scroll. The free Adventures mode was an honest attempt to re-impose a shape on that pile, and it half works.

What it is actually for

Here is the read.Vampire Survivors is a game about the pleasure of a system you built becoming legible to you all at once. The reason it survives its own tedium — and there is tedium; minutes twenty-two to thirty of a strong run are frequently just holding a direction — is that the legibility arrives as asensation rather than as information. You do not read a stat sheet and conclude the build worked. You watch the screen fill with your own consequences.

That is a real, specific thing that games can do and other media cannot, and it is why the hundred clones that copied the auto-attack and the XP gems mostly feel hollow. They copied the loop. The loop was never the point. The point was the ten-minute window of genuine fear that makes the twentieth minute mean something, and fear is expensive to design and free to leave out.

Play it on whatever you have — it is on PC, Xbox and Game Pass, Switch, PlayStation, and free on mobile with ads, and it runs on hardware that would struggle to open a browser tab. Buy the base game, ignore the DLC until the base game bores you, and give it the first hour on the understanding that the first hour is meant to be a bit miserable.

If it hooks you, go toRisk of Rain 2 next, which does something structurally related with a clock that gets angrier the longer you dawdle, and which asks for rather more of your hands.

Spoilers below

The unlock tree is the real second game, and it is where the design shows most personality.

Moongolow is the trick everyone remembers: a stage that runs a fixed fourteen minutes and then ends in a cataclysm, dropping you somewhere considerably less pleasant, and it is the only moment inVampire Survivors that has anything you could call a set piece. It works because the game has spent hours teaching you that stages are inert arenas, and then one of them turns out to have a plot.

Green Acres and The Bone Zone are jokes — endless, arbitrary, essentially test rooms left in with the lights on — and their presence tells you something honest about the project’s origins. This was a browser toy that never fully stopped being one, and Poncle has declined to sand off the parts that give it away. The secret characters unlocked by typing nonsense into the main menu are the same instinct: a game that remembers cheat codes were once a folk practice rather than a store page.

The Randomazzo, the Yellow Sign, the escalating chain of relics that each unlock the ability to find the next thing — this is the structure that keeps people at four hundred hours, and it is essentially an ARG bolted to a screensaver. Whether that is genius or a cheerful mess depends entirely on how you feel about a game that hides its best ideas behind its worst ones. I lean towards genius, on the grounds that nothing else that came out in 2022 was this confident about what it could afford to leave out.

Roguelikes have always had a story problem, and it is a structural one. The genre runs on repetition — you die, you start again, the numbers reshuffle — and narrative runs on progression. Put the two in the same box and the story becomes something the player skips. Every roguelike before 2020 solved this by having almost no story, or by putting it in item descriptions and letting the community assemble it on a wiki.

Hades solved it by making the repetition the subject.

Supergiant Games put it into early access on the Epic Games Store in December 2018, moved it to Steam a year later, and shipped 1.0 on 17 September 2020 for PC and Switch, with PlayStation and Xbox versions following in August 2021. It won a BAFTA for Best Game and, in 2021, the Hugo Award in a one-off Best Video Game category — the first game to get one, voted on by a science-fiction readership that does not hand those out for combat feel. Five years and a sequel later, the thing it worked out about narrative repetition is still the most important design idea of its generation, and it is still barely copied, because copying it is enormously expensive.

Death is a commute

Zagreus is the son of Hades. He is trying to leave the Underworld. When he dies, he goes into the River Styx and surfaces in a pool in the House of Hades, which is his father’s office, and his father looks up from his paperwork and says something about it.

That is the entire trick, and everything else follows from it. Once dying returns you to aplace where people live, the run structure stops being a loop and becomes a commute. You leave home, you fail, you come home, and everyone at home has an opinion about your failure. Achilles is by the door. Nyx is in the hall. Cerberus wants attention. Dusa is dusting the chandelier and worrying. Hypnos — who is asleep at the reception desk and reads out your cause of death like a receptionist reading a delivery note — is the joke that makes the whole thing work, because he turns each death into an event the fiction acknowledges.

CompareRogue Legacy (Cellar Door Games, 2013), the closest ancestor, which made death diegetic first: your heir inherits the estate and the traits, and the castle persists. That is the right idea, executed as a frame.Hades runs the idea through the writing, and the writing is the part nobody wants to pay for.

The dialogue queue is the actual engine

Here is the machinery, and it is worth understanding because it is the whole answer.

Creative director Greg Kasavin wrote north of twenty thousand lines of dialogue forHades, all of it voiced, most of it by a small cast with Logan Cunningham carrying an implausible share of it. Plenty of RPGs have more words than that, so the volume is only half of it. The innovation is the priority system underneath.

Every character has a large pool of possible lines, each tagged with conditions: what boons you carried, who you last spoke to, which boss killed you, how many runs you have made, what you gave whom, what you have already been told. When you walk up to Achilles, the game queues the most contextually specific line that has not yet fired, and burns it. Say the wrong thing at the wrong time and the game notices; die to your father twice in a row and he has a fresh remark about it.

The player-facing consequence is thatHades almost never repeats itself for the first forty or fifty runs, and by the time it starts to, you are deep enough that the story has moved. The illusion is that the House is reacting to you. The reality is a very large deck being dealt in an intelligent order, and it holds because Supergiant did the unglamorous work of writing enough cards.

That is why nobody has copied it. The mechanic is cheap. The content pipeline feeding it is not.

The run itself

None of this would matter if the combat were poor, and it is excellent for a reason that has nothing to do with the writing: the boon system creates a build, and the build is a conversation with chance.

Six weapons, each with aspects that alter them substantially. Boons from the Olympians — Zeus chains lightning, Poseidon knocks back, Aphrodite weakens, Ares does damage over time, Artemis crits, Dionysus poisons — and boons that combine into Duo effects when the right two gods have already blessed you. The Mirror of Night spends darkness on permanent upgrades. Keepsakes weight the pool towards a god you want. Chthonic keys, nectar, ambrosia, the Fated List of Minor Prophecies: every currency you bring home buys something.

The design pressure this creates is genuinely good. You cannot plan a build. You can only lean — take Artemis’s keepsake, hope she shows up, adapt when she doesn’t. It is the same tension I wrote about inBalatro: The Poker Roguelike That Ate a Year: the run gives you a hand and the skill is recognising what hand you have been given rather than the one you wanted.Hades is more forgiving thanDead Cells about this, because the meta-progression means a bad run still pays, and that forgiveness is deliberate — the design wants you to get home so the House can talk to you.

The accessibility work belongs in the same argument. God Mode grants 20% damage resistance and adds 2% every time you die, so a player who keeps failing keeps getting stronger until the story unlocks. Supergiant understood that the story was the reward and refused to gate it behind execution. The Pact of Punishment runs the other way for players who want the difficulty back, in increments they choose.

What it owes

The Amiga I got in 1987 had a port ofRogue on it — Epyx published one in 1986 — and the thing aboutRogue is that it had no story at all and did not need one, because the dungeon generated the anecdote. You told the story afterwards, in a corridor at school. That is the genre’s founding compromise: the game supplies systems and the player supplies meaning.

Every roguelike since has honoured that compromise.Spelunky (2008),The Binding of Isaac (2011),FTL (2012) — all of them make you the author. What Supergiant did was ask whether a roguelike could supply the meaning itself without losing the generative anecdote, and the answer turned out to be yes, at a cost most studios cannot bear. Supergiant had also solved half of it already: the reactive narrator inBastion (2011) was the same technology in an earlier and cruder form, a system watching what you did and commenting on it.Hades is that prototype, nine years of iteration later, pointed at the exact structural problem it was built to solve.

The verdict

Hades is the rare game where the literary ambition and the mechanical design are the same object. Take the writing out and you have a very good isometric action roguelike with a strong art direction from Jen Zee and one of Darren Korb’s best scores. Take the combat out and you have a soap opera about a dysfunctional family of gods. Together they produce something neither half could: a story that getsbetter the more you fail at the game, which no other medium can do at all.

Its limits are honest ones. The Underworld’s four biomes are fixed in order — Tartarus, Asphodel, Elysium, the Temple of Styx — so the variation lives in the boons and the room layouts, and after enough runs the geography is a hallway you walk fast. The bosses are few, and Theseus and Asterius carry more than their share. The moment-to-moment combat lacks the mechanical strangeness of the best of its peers.

It is on everything — PC, PlayStation, Xbox, Switch — it runs on a toaster, and it is a complete game that never asked for a season pass.Hades II followed the same early-access route and only widens the argument: Supergiant found a design position nobody else can afford, and they are still the only ones standing on it.

What to play next:Dead Cells: The Roguevania Blueprint for the harsher, more mechanically dense version of the same loop, andRisk of Rain 2: The Difficulty Curve as a Clock for a roguelike that solves pacing with time pressure instead of dialogue.

Spoilers below

The ending is the part people argue about, and the argument is a good one.

Zagreus escapes, reaches the surface, finds Persephone, and then dissolves — because he cannot survive outside the Underworld, and the game makes you do this repeatedly. Ten times, in fact, before the story concludes. Supergiant took the one thing a roguelike player wants (the winning run) and made it a chore you must grind, which sounds like a design failure and is instead the sharpest joke in the game: reaching Persephone is the beginning of the story, and the reunion has to be earned through the same repetition everything else was.

The epilogue lands the thesis. Persephone comes home, the family is assembled, and the whole thing closes on a family reconciling — Hades and Zagreus finally speaking plainly, Nyx’s role revealed, Demeter thawing. It is a domestic ending to a myth about escape, and the reason it works is that you have spent sixty hours in that hallway hearing these people slowly change their minds about each other, one death at a time.

Hypnos, who has been reading out your causes of death for the entire game with no idea what is going on, gets the last laugh. Correctly.

The first hour ofControl is a woman walking into an office. There is a janitor with a Finnish accent, a reception desk, a directory board, and a pyramid of concrete overhead that goes up further than the building’s exterior allows. Remedy Entertainment released it on 27 August 2019 — PC via the Epic Games Store for its first year, plus PS4 and Xbox One — under director Mikael Kasurinen with Sam Lake writing alongside Anna Megill and Brooke Maggs, and published by 505 Games. Six years on it is the studio’s most complete piece of world-building, and the reason is architectural rather than narrative.

The Oldest House is the best level in any game of its decade, and it is a level about bureaucracy.

The building is the design document

Brutalism is a real aesthetic argument: the structure is the ornament, the concrete is left showing, the building declares its own systems. Remedy took that literally. The Federal Bureau of Control occupies a mid-century concrete government block, all board-marked walls, terrazzo floors, wood panelling, green-shaded lamps and an internal mail system, and then the building starts behaving like an Object of Power. Corridors reconfigure. A stairwell delivers you somewhere geometry says it should not. The Ashtray Maze rearranges itself faster than you can walk it.

This is a very specific horror, and it is the reason the game sticks. Haunted mansions are exhausted; a hauntedadministrative facility is not. The dread inControl comes from paperwork — a redacted case file describing a rubber duck that killed six people, a research memo written in the flat voice of a civil servant who has stopped being surprised. The building’s uncanniness is legible only because everything around it is so aggressively ordinary. You believe the impossible pyramid because the noticeboard next to it has a poster about workplace ergonomics.

The lineage is legible and Remedy has never hidden it: the SCP Foundation’s clinical containment prose, Mark Z. Danielewski’sHouse of Leaves for the building that is larger inside, Jeff VanderMeer’sAnnihilation for the New Weird tone,Twin Peaks for the FMV inserts and the Finnish rock band. What Remedy added is the thing prose cannot do: youwalk it. A house that is bigger on the inside is a conceit on the page and a spatial fact in an engine.

Why the combat holds

The Service Weapon is the second-best idea in the game. It is a single pistol that reconfigures into five forms — Grip, Shatter, Spin, Pierce, Charge — with no ammunition, only a recharge, so the weapon is a tool you select rather than a resource you manage. Beside it sits Launch: telekinesis, on a cooldown, which tears a chunk of the building loose and throws it at somebody.

The reason this feels good is Northlight, Remedy’s engine, and specifically the destruction. The Oldest House is built out of debris waiting to happen. Every desk, chair, filing cabinet, monitor, partition wall and potted plant is a projectile, and a serious fight in the Bureau’s open-plan offices ends with the room reduced to particulate. The combat loop therefore runs: shoot to build energy, Launch to spend it, watch the office disassemble. Levitate, added later in the ability tree, lifts the whole thing into three dimensions and turns the atria into arenas with a Z-axis.

The real ancestor here sits outside Remedy’s own catalogue. It isPsi-Ops: The Mindgate Conspiracy (Midway, 2004), which built a whole third-person shooter around telekinesis and physics objects and which nobody bought, alongsideHalf-Life 2’s gravity gun (2004), which taught a generation that a physics object in your hands is more interesting than a bullet. Both games arrived when physics middleware was new enough to be the selling point, and both understood the same thing: the pleasure of throwing a filing cabinet is that the cabinet was furniture a second ago.Control is the first game to give that idea a budget, an art director and a building worth destroying.

Where it fights itself: the enemy variety is thin. The Hiss are men in hard hats and body armour, floating, chanting, and by hour twelve you have seen the roster. The encounters escalate through numbers and health pools, which is the least interesting axis available to a game with this much physics under it.

The ability tree compounds it. Jesse’s powers arrive on a schedule tied to Objects of Power, and each is excellent on arrival, yet the game rarely builds an encounter that demands two of them together. Seize — turning an enemy to your side — is the clearest waste: a mechanic with real tactical depth, deployed against enemies who die too quickly for the investment to matter. Remedy built a sandbox and then mostly asked you to clear rooms in it.

The map is a genuine failure

I take the map seriously becauseControl takes navigation seriously and then sabotages it.

The Oldest House is a Metroidvania — gated sectors, clearance levels, ability locks, backtracking — and it ships with a map that does not rotate, does not sensibly express vertical relationships, and is close to unreadable in the multi-level sectors it most needs to explain. In a game whose entire subject is a building that will not hold still, being lost is thematically perfect and practically miserable. CompareMetroid Dread, where the map is the interface the whole game is played through.

The launch checkpointing was worse and Remedy fixed it. Control Points were sparse, and a boss death could send you on a long walk back through cleared rooms; the studio patched in additional checkpoints — including around the Anchor fight and the mould sequence — after the complaints landed. Base-console performance was rough at launch too, and theUltimate Edition in August 2020 brought ray tracing to PS5 and Xbox Series consoles, though existing owners on older consoles found the upgrade path handled badly enough to become its own small scandal.

The verdict

Control is a game whose ideas outrun its systems, and it is worth playing for the ideas. The Oldest House is a genuine achievement of environmental design — the rare fictional space you could navigate in your head years later — and the Bureau’s tone, that mixture of cosmic horror and departmental procedure, is Remedy’s most original register. The physics give the combat a texture nothing else quite matches, even when the enemies opposing you are dull.

What it does not have is a second act with the confidence of its first. The Hiss are a weak antagonist for a building this strange, and the mid-game settles into a rhythm of side missions that ask you to clear a room with the same three enemy types you cleared the last room with. The Ashtray Maze is the correction — a scripted, musical, twelve-minute sequence where the level design and the soundtrack take over completely — and its presence in the last quarter is a reminder of how much more the building had left in it.

The Foundation (March 2020) and AWE (August 2020) expansions are worth taking. The second folds Alan Wake into the Bureau’s case files and turns Remedy’s back catalogue into a shared universe, which they then cashed in properly withAlan Wake 2: Remedy’s Swing at the Fence. Remedy bought the fullControl rights back from 505 Games in 2023 and a sequel is in development, which is the correct outcome for a studio that finally built a world worth owning.

Play theUltimate Edition on PC or a current console. It is cheap, it is everywhere, and there is nothing else like the Oldest House.

What to play next:Alan Wake 2 for what Remedy did once it stopped apologising for being strange, andSignalis: The Survival Horror That Reads Like a Poem for bureaucratic dread on a fraction of the budget and twice the compression.

Spoilers below

The Ashtray Maze is the game’s high point and it is instructive to say why, because the sequence works by taking things away. For twelve minutesControl removes navigation entirely — the maze folds and unfolds itself, you cannot get lost because there is no choice to make — puts Poets of the Fall’s Old Gods of Asgard over the top of it, and hands you a corridor of fights to walk through while the building performs for you. Everything the game has been fumbling (pacing, legibility, the sense that the Oldest House is doing this deliberately) snaps into focus the moment the player’s agency is narrowed.

Which raises the awkward question:Control’s best sequence is the one where it behaves least likeControl.

Jesse’s internal monologue — the second voice she has been talking to since childhood, revealed as Polaris, a benign Object of Power riding along — is the structural gag that pays off the FMV inserts and Dr Darling’s increasingly unhinged research films. Trench’s suicide, delivered as the game’s opening beat and understood only later as a man refusing to become a vector for the Hiss, is the sharpest piece of writing Sam Lake has done. And Dylan, the brother, is the Bureau’s real indictment: an agency that studies children as containment risks and calls the paperwork ethics.

The ending withholds resolution deliberately — Jesse is Director, the Hiss are contained rather than defeated, the building keeps its secrets. That was a sequel hook in 2019 and it looks like patience in 2025.

Titanfall 2 came out on 28 October 2016.Battlefield 1 had arrived on 21 October.Call of Duty: Infinite Warfare landed on 4 November. Respawn’s game was released by EA into a fourteen-day gap between EA’s own military shooter and the biggest annual shooter on earth, and it was flattened. Patrick Söderlund, then running EA Studios, said publicly the following year that the timing had hurt the game and that they had misjudged it. That is the whole tragedy in one paragraph, and it is the reason the game is now shorthand for a certain kind of industry injustice.

Nine years on, the injustice framing has calcified into a meme, which does the game a disservice.Titanfall 2’s campaign survives on its own merits: it does something structurally unusual that almost nobody has copied since, and the reason nobody has copied it is that it is expensive, wasteful, and hard. The bad luck is a footnote to a design worth studying.

One mechanic per level, then bin it

The standard shooter campaign teaches you a verb in hour one and spends the next seven hours escalating it. More enemies, bigger rooms, a boss.Titanfall 2 does the opposite: each level introduces a mechanic, extracts a single good idea from it, and throws it away before you get bored.

You get a level where the terrain itself is the assembly line: a factory that builds rooms in front of you as you move through them, so the level is being constructed at the speed you traverse it. You get a level built around a time-shift device that flips you between two eras of the same facility with a button press, where the enemies, the geometry and the light all change and your momentum does not. You get a gauntlet — a speed-run training course with a leaderboard, sitting inside a campaign, teaching you your own movement by timing it.

Respawn has been open about how this came about: the team built playable prototypes first, small self-contained toys, and then designed levels around the ones that felt good. The time-shift level, “Effect and Cause”, came out of that process under designer Mohammad Alavi. The result is a campaign that reads like a compilation album, and it is short — five or six hours — precisely because a mechanic gets one level and then leaves.

This is a real design position, and it costs money. Every one of those toys is bespoke engineering for a single level, thrown away afterwards. It is why the game has no filler and also why nobody in a boardroom wants to greenlight it. A campaign built this way cannot amortise its costs across a sequel. CompareDoom Eternal: The Shooter as Puzzle, where id built one economy and spent eighteen hours proving it. Both are disciplined. Only one is repeatable.

Why the movement is the actual argument

The mechanics-per-level structure gets the headlines. The thing underneath it is better.

Pilot movement inTitanfall 2 is a wall-run, a double jump, and a slide that preserves speed, and the crucial property is that none of them are on a cooldown or a meter. Chaining them is a matter of the level agreeing to be chained. So every space in the game is authored twice — once for a player walking through it and once for a player who never touches the floor — and the design’s whole character comes from that second reading.

What this produces is a feeling almost nothing else in the genre offers: the levels are legibleas movement. You look at a room and you see a route rather than a set of cover positions. When the time-shift level asks you to jump between eras mid-wall-run, the reason it works is that both eras were built to be run through, so the switch never breaks the line you are drawing through the space.

The Titan half is the counterweight, and it is smarter than it looks. BT-7274 is slow, heavy, and armed with loadouts you swap by taking them off dead Titans, and the deliberate friction after twenty minutes of pilot movement is what makes both halves read. A game that was all wall-running would flatten into noise inside an hour. The campaign alternates the two registers relentlessly — light and fast, then heavy and considered — and the rhythm is the reason six hours never sags.

There is a design lesson here that the industry mostly ignored: the pleasure of mobility depends on periodically taking it away.

The one that got away

Titanfall 2’s time-shift level shipped two weeks beforeDishonored 2 (11 November 2016), which contains “A Crack in the Slab”, a level built on the same idea — a device that flips you between two time periods of the same building. Two studios, parallel development, no possibility of copying, both landing the trick in a fortnight of each other. It is one of the strangest coincidences in level-design history, and the fact that both are widely considered their game’s best level suggests the idea was simply sitting there waiting for hardware that could hold two versions of a space in memory at once.

The genuine ancestor of “one bespoke mechanic per chapter” is older:Half-Life (Valve, 1998), which introduced a set piece, resolved it, and moved on, and whose sequel built a physics toy and then spent a chapter on the gravity gun before dropping it. Respawn’s team came out of Infinity Ward, which came out of 2015 Inc., which madeMedal of Honor: Allied Assault — a lineage that has been building scripted, single-use spectacle since 2002.Titanfall 2 is what that tradition looks like when the spectacle is handed to the player as a system rather than played at them as a cutscene.

What went wrong, and what happened after

The release window is the famous part. The less famous part is the aftermath. Respawn’s multiplayer — free maps for everyone, no season pass carving up the population — was the most player-friendly network model any big publisher shipped that year, and it was undermined for years by sustained attacks on the servers that left matchmaking unreliable or unusable for long stretches. A community project, Northstar, eventually stood up unofficial servers so people could play. EA rolled out fixes in 2024 that restored official matchmaking. That an eight- year-old shooter needed rescuing twice, once by its players and once by its publisher, is a fair summary of how the game has been treated.

EA acquired Respawn in 2017. The studio then madeApex Legends, set in the same universe, released it in February 2019 with no announcement, and it became one of the biggest games in the world. Every wall-run and double jump inTitanfall 2 is money that eventually arrived, just wearing a different hat.

The verdict

The campaign is six hours long and has no fat on it whatsoever, and that is a genuinely rare sentence to be able to write about a AAA shooter. The mechanics are disposable by design, the movement is the best in the genre, the pacing between pilot and Titan is exact, and the whole thing is over before it can disappoint you. Its reputation as the great lost shooter is deserved on the merits and has nothing to do with the release date.

The reservations are real. The story is thin — competent, warm, thin — and the emotional weight it goes for in the last hour is doing a lot of work on very little setup. The Titan boss fights against the Apex Predators are the least interesting encounters in the game, arriving on a schedule and reading as a different, more ordinary shooter’s idea of structure. And six hours is six hours; the campaign gives you a movement system that only becomes properly expressive around hour four.

It is on PC, PlayStation and Xbox, it is old enough to be permanently cheap, and it runs on everything. The campaign alone justifies it.

What to play next:Doom Eternal: The Shooter as Puzzle for the opposite structural bet — one system, endlessly deepened — andMetroid Dread: The Series Remembers What It Is for another design that treats mobility as the thing being authored rather than the way you get to the authoring.

Spoilers below

“Effect and Cause” is the level everyone names, and it deserves it, but the better piece of design is “Into the Abyss” — the factory that assembles the level in front of you. It is the game’s thesis made physical: the space is being authored at the speed you move, and you are outrunning your own level designer.

BT’s ending works despite the setup being thin, and the reason is mechanical rather than narrative. You have spent six hours climbing into and out of BT, being caught by BT, being thrown by BT across gaps you could not cross alone. The throw is a verb you have executed dozens of times. When the last one comes, the game asks you to perform an action you have internalised, one final time, with a different meaning attached — the feeling arrives through your hands, having been rehearsed for six hours under another name. That is what games can do that film cannot, and Respawn got there through a mechanic they had been quietly teaching since level two.

The Ark and the Fold Weapon plot is disposable. Nobody remembers it. Everybody remembers the throw.

I playedDoom in 1993 on a beige tower that a friend’s older brother had talked his parents into, and what I remember is the width of it — you strafed, you circled, you emptied a shotgun into a room and the room emptied back. The game asked for nerve and spatial sense. It did not ask for a plan.

Doom Eternal asks for a plan. id Software released it on 20 March 2020 for PC, PS4, Xbox One and Stadia, with a Switch port arriving that December, and it remains the most divisive big-budget action design of its generation for one reason: game director Hugo Martin and his team took a shooter and built a puzzle game inside it. Five years on, the people who bounced off it have not come back, and the people who clicked with it have never really stopped playing. Both camps are responding to the same mechanic. They just disagree about whether a first-person shooter is allowed to have a correct answer.

The triangle

Here is the design, stated as plainly as it deserves.

Ammunition is scarce, and the reserves are small enough that you will empty a weapon during any serious fight. Your refill is the chainsaw, which regenerates one fuel pip on a timer, and one pip is enough to open a small demon and spray ammo across the floor.

Health is scarce, and the pickups are thin. Your refill is the Glory Kill — stagger a demon, punch it apart, take the health it drops.

Armour is scarce. Your refill is the Flame Belch, which sets a demon burning so that everything you subsequently land on it sheds armour shards.

Three needs, three verbs, each verb requiring you to be inside the fight rather than backing out of it. That is the whole engine, and it is whyEternal moves the way it does. The 2016 reboot introduced the Glory Kill and called the philosophy push-forward combat.Eternal takes the same idea and closes every exit. You cannot turtle, because turtling starves you. You cannot hoard, because the reserves are too small to hoard into. The only route to resources runs through the demon standing in front of you, and the game has arranged for a demon to always be standing in front of you.

Why the arena is a lock

The second layer is where the argument really lives: weapon-specific weaknesses.

The Mancubus has arm cannons that come off to a precise shot. The Revenant has shoulder launchers. The Arachnotron has a turret on its back that can be sheared away, which stops it suffering you at range. The Cacodemon will swallow a grenade fed into its open mouth and go straight into a stagger. The Carcass throws up a barrier that the Super Shotgun tears down. The Pain Elemental floats where only certain tools reach.

Every demon, then, is a lock with a named key, and an arena is a queue of locks opening at you simultaneously. The mental loop that results is genuinely strange for a shooter: the dominant activity issorting. Which threat resolves fastest, which weapon is loaded, do I have the fuel to open a fodder demon for ammo before the Mancubus commits, can I dash twice to reposition before the Revenant’s second volley lands. The pace is frantic and the thinking is turn-based. Hugo Martin has described the fantasy as being an apex predator running a chess board at 200 miles an hour, and that is an accurate description of what the pad is doing.

The result is a combat system where two competent players can look completely different. One is playingDoom Eternal. The other is playingDoom (2016) inDoom Eternal’s costume, running out of ammo every fifteen seconds, and hating every minute.

Why people hate it

The complaint is coherent, and dismissing it is lazy.Eternal removes freedom from a series whose entire cultural memory is freedom. If you want to solve a room with the rocket launcher because the rocket launcher is fun, the game will punish you for it, and the punishment arrives as an empty magazine with a Mancubus attached. The design has an opinion about how you should play and it enforces that opinion through the economy rather than through a difficulty slider. Some players experience that as being taught. Others experience it as being managed.

The Marauder is where the argument gets loudest, and it is the fairest test case. He blocks everything outside a specific range band, and he opens for exactly the window in which his eyes flash green — a window you cannot rush, cannot bait early, and cannot skip. He is a rhythm-game boss dropped into a shooter. Fight him correctly and he is the best encounter id has ever built; fight him the way you have fought everything else since 1993 and he is a wall that laughs at you. He is a purer version of whatSekiro does to people, and he provokes the same split: some players hear the game telling them the tempo, and some hear it telling them off.

The platforming is a weaker defence. The monkey bars and the jump pads exist to give the arenas breathing room, and they largely do, but they also drag a game whose best quality is momentum into stretches where the momentum is a chore. Nothing in the traversal is as interesting as the worst fight.

What it is actually descended from

The easy read is thatEternal descends fromDoom II (1994), and it does in the sense that combined-arms encounter design — an archvile behind a crowd, a revenant on a ledge — is where id first learned to make a room think. The deeper ancestor is elsewhere.

Eternal’s real lineage runs through the resource-economy roguelike. The question the game asks you every four seconds — what do I spend, what do I have left, what do I need to be holding twenty seconds from now — is the questionRisk of Rain 2 asks with its clock, which I wrote about inRisk of Rain 2: The Difficulty Curve as a Clock. It is also the question every good arena-shooter map asked in 1999, when the armour and the quad respawned on timers and the entire skill of Quake was knowing where the resources would be before they existed. id’s insight in 2020 was to move that timer off the map and onto the demons. Once the ammo is inside the enemy, spatial control and resource control become the same act, and the shooter collapses into a single verb.

For the boss-rush reading of a similar idea — encounters as locks, each with a tool — the mech that best rhymes with it is inArmored Core VI: The Boss Rush Hiding in a Mech Game.

The verdict

Doom Eternal is the most rigorously designed action game of its era, and the rigour is the reason it will never be universally liked. It is a game with a thesis, and it spends eighteen hours proving the thesis at you. When the loop locks in — chainsaw, Glory Kill, Belch, Super Shotgun into a Cacodemon’s mouth, meat hook to the next problem — nothing else in the genre produces that specific feeling of a hard system dissolving into instinct. When it does not lock in, you are standing in a car park with no bullets, and the game feels like homework someone set you.

The post-launch record backs the thesis.The Ancient Gods, Part One (October 2020) andPart Two (March 2021) tightened the screws rather than loosening them, adding encounters aimed at players who had already internalised the economy. Update 6 added Horde Mode in 2021 — pure arenas, no story, no monkey bars — which is id quietly agreeing about what the game is for. The Denuvo Anti-Cheat component added in May 2020 was pulled within days after players objected, a small episode that says more about the PC community than the game.

Play it on anything that will run it at a high frame rate; the design assumes your inputs are instant and it becomes a different, worse game when they are not. If the economy irritates you in the first two hours, it will irritate you in the last two. That is a real answer, and it is not a failure of the game.

What to play next:Risk of Rain 2 for the same resource pressure expressed as time, andSekiro: The Rhythm Game With a Sword for the other 2019–20 combat system that insists there is a correct answer and refuses to accept anything else.

Spoilers below

The campaign’s structure is the weakest thing about it, and the lore is whereEternal most conspicuously loses its nerve. The 2016 reboot understood that the Slayer worked because he was a blank — a man who threw a monitor across a room rather than listen to exposition.Eternal answers questions nobody had: the Maykrs, Urdak, Argent D’Nur, the Slayer’s origin as Doom Guy pulled out of a Night Sentinel order, the whole Father-and-Khan-Maykr theology. The joke of 2016 was that the story was an obstacle the protagonist wanted removed. Making the protagonist the centre of a cosmology retires the joke.

The Ancient Gods, Part Two closes it out with a duel against the Dark Lord, a one-on-one that strips away the arena entirely and asks you to parry — the final statement of the Marauder’s argument, aimed directly at the players who spent two years insisting the Marauder was badly designed. It is a fight only somebody who had accepted the game’s terms could enjoy, and it is the most honest endingEternal could have had.

On 25 January 2023, Xbox ran a Developer_Direct, announcedHi-Fi Rush, and put it on sale the same hour. No marketing run-up, no preview embargo, no two-year drip of trailers. Tango Gameworks — the studio Shinji Mikami founded in 2010, best known at that point forThe Evil Within andGhostwire: Tokyo — had made a cel-shaded rhythm-action comedy under director John Johanas and simply let it go. It is the cleanest shadow-drop of the decade, and it worked precisely because the game does something you cannot really convey in a trailer. You have to hold the pad.

Two and a half years on, the release-day novelty has burned off and the design is still standing up, which is the only test that matters. What is left is the most interesting argument anyone has made about rhythm in an action game sinceRez.

Everything is on the beat, including the furniture

The core fact:Hi-Fi Rush runs at 120 beats per minute, and it runseverything at 120 beats per minute. Chai’s attacks resolve on the beat. Enemy telegraphs land on the beat. Platforms rise and fall on the beat. Fans turn, hoardings flash, girders swing, and a robot cat called 808 wags its tail — all on the beat. The world is the click track, rendered.

This is the bit that people underrate. Most rhythm games put the music in one channel and ask you to match it with your thumbs.Hi-Fi Rush dissolves the channel. Every readable object in the frame is a metronome, so you are never listening for the beat as a separate task — you are reading the room, and the room happens to be in 4/4. Losing the beat here feels like losing the plot of a conversation: you fall out of sync with a place, and the place tells you so from every corner of the frame at once.

The fiction earns it, too. Chai is a wannabe rock star who volunteers for Vandelay Technologies’ Project Armstrong to get a prosthetic arm, and his music player falls into the machinery, fuses to his chest, and makes him a defect: the one person in the building who can hear the rhythm the building is moving to. That is the diegetic justification for a HUD that would otherwise be a bar at the bottom of the screen — the beat becomes a fact about Chai before it becomes an interface element.

The generosity is the mechanic

Here is where the design gets genuinely clever, and where it departs from thirty years of rhythm-game orthodoxy.

You cannot fail for being off-beat.

Press attack at the wrong moment and Chai still swings — the game quantises the input to the next beat and lands it. Nothing punishes you. Nothing shatters, no combo counter resets to zero out of spite, no “MISS” strobes across the middle of the fight. What you lose isupside: on-beat hits do more damage, extend combos further, and pay out better ratings at the end of the encounter. The beat is a multiplier on top of a competent character-action game rather than a gate in front of it.

Compare that toCrypt of the NecroDancer (Brace Yourself Games, 2015), which is the closest structural ancestor and which takes exactly the opposite line: step off the beat and your gold multiplier collapses, and in the harder modes your turn simply does not happen. That is a purist’s design and I like it, but it sorts players into those with rhythm and those without within ninety seconds. Tango’s version sorts nobody. It lets a player with no sense of time at all finish the game while still making the beat feel like the most interesting thing in it, because the feedback for finding it is so lavish — the hit pauses, the partner assist chimes in on the downbeat, the whole encounter suddenly reads as choreography instead of as work.

That generosity extends outward. There is a beat-visualiser you can leave on permanently. There is an accessibility option that auto-aligns your inputs outright. The rhythm parry — which arrives partway through, taught by Korsica after she stops trying to kill you — is a timing window like any parry, and the game gives you a metronome to hit it with.Sekiro asks you to internalise the tempo of a duel with no click track at all; I wrote about why that works inSekiro: The Rhythm Game With a Sword, and the interesting thing is thatHi-Fi Rush arrives at a comparable feeling of locked-in flow by handing you the very information FromSoftware withholds. Both work. They just disagree about whether the tempo should be a secret.

What it owes, and to whom

The visual lineage is legible from the first corridor: the cel shading and the comic-panel sound effects come out ofJet Set Radio (Smilebit, 2000) andViewtiful Joe (Clover, 2003), and the combat vocabulary — light and heavy strings, launchers, a rating screen that grades you at the end of every scrap — is Capcom’sDevil May Cry school, filtered through people who clearly enjoy it without wanting to be brutalised by it.

The real ancestor of thefeeling, though, is older and further sideways. In 1987 I had an Amiga, and the thing that machine did better than anything else in the room was sync visuals to a tracker module — a demo where the bassline drove the geometry, where the music caused the scene.Hi-Fi Rush is that trick, thirty-five years on, with a budget and a character-action game hung off it. Tetsuya Mizuguchi’sRez (2001) is the other obvious forebear, and the connective tissue between all of them is the same insight: the pleasure lies in beinginside a system where causation runs through the rhythm, so that hitting the beat feels like agreement rather than obedience.

Where it fights itself

Three real problems, in ascending order of how much they cost.

The comedy is loud and it is relentless, and it is aimed squarely at a young audience. Chai is written as an idiot with a good heart, and the jokes come at you in a stream. It landed for me more often than it missed. Anyone allergic to the cadence will find eight or nine hours of it a long time.

The rating pressure sits slightly askew from the generosity elsewhere. The game tells you the beat does not matter, then grades your entire encounter on how well you kept it. That tension is productive for most of the run and merely nagging in the back half, when the bosses start demanding rhythm-parry sequences that tighten the window the design has spent hours telling you is loose.

And the platforming is the padding. The traversal between fights is a rhythm game in the least interesting sense — hit the jump on the beat, land on the moving thing — and it exists to space the combat out. Tango knows the combat is the product; the level design keeps interrupting it to prove the world is on the beat, which the world had already proved.

The post-launch work argues the studio knew where the value was. The freeArcade Challenge! update later in 2023 addedPower Up! Tower Up!, a run-based mode with randomised modifiers, andBPM Rush!!, which ratchets the tempo upward as you go. Both do the same thing: strip out the platforming and give you the fights.

The verdict

Hi-Fi Rush is the rare game whose central conceit is load-bearing. The metronome is the skeleton, and the fighting, the level readability and the tone all hang off it correctly. It is the best argument I know that rhythm mechanics do not require punishment to produce flow, and that a design can hand a player every piece of information and still leave them something to master.

Its afterlife is the ugly part. Microsoft closed Tango Gameworks in May 2024, months after the studio shipped a well-reviewed original game and an award-season favourite. Krafton subsequently acquired the studio and theHi-Fi Rush IP in a deal announced that August. The game came to PS5 in March 2024. It is on PC and Xbox, it runs on anything, and it will still be doing this trick in ten years, because 120 BPM does not go out of date.

What to play next:Sekiro: The Rhythm Game With a Sword for the same flow state with the click track removed, andNine Sols: The Sekiro Parry in a Taopunk Frame for a 2D parry loop that also believes tempo is the whole conversation. If you want the purist counter-argument to Tango’s kindness,Crypt of the NecroDancer is right there, waiting to fail you.

Spoilers below

The structural joke of the back half is that Vandelay’s executives are each a department made flesh — Rekka in HR, Zanzo in R&D, Korsica running security, Mimosa in QA, Roquefort selling — and each boss fight is a satire of what that department does to a person. The QA fight is the best of them, because the encountertests you the way QA tests a build, and the game is self-aware enough to make the joke land mechanically rather than only in dialogue.

Korsica’s defection is the pivot that fixes the combat. Up to that point Chai is a fairly conventional combo machine with an assist button. The rhythm parry arrives, and every subsequent encounter is a two-way conversation with a tempo rather than a one-way beating. It is the correct place to put it — late enough that you have internalised 120 BPM, early enough that you get to enjoy it.

And the Project Armstrong reveal — that Kale’s rollout is a plan to remove the defective from the workforce entirely — is a broader corporate satire than the first hour prepares you for. The game keeps its silly voice while making its point, which is harder than it looks.

Metroid Dread was announced in 2005 for the Nintendo DS, went quiet, was rumoured for a decade and a half, and then turned up in October 2021 on Switch as if the intervening sixteen years had been a scheduling error. Produced by Yoshio Sakamoto and developed by MercurySteam with Nintendo EPD, it is Metroid 5 — the direct sequel toFusion, which came out on the Game Boy Advance in 2002. Nineteen years between instalments of a numbered series is a long time to think about what you are.

The remarkable thing is thatDread clearly used the time. This is a game with a very specific and slightly severe opinion about what a 2D Metroid is for, and the opinion isn’t the one most of the genre’s descendants arrived at. The rest of the metroidvania field spent twenty years adding — skill trees, roguelike runs, dialogue, builds.Dread subtracts. It plays like somebody re-readSuper Metroid, wrote down the three things that made it work, and threw the rest overboard.

Movement is the whole subject

PlayDread for ten minutes and the first thing that registers is how fast Samus is. She slides. She free-aims in any direction while walking. She has Flash Shift — a short teleporting dash, chainable — from early in the run. The Speed Booster is still here, the Shinespark still fires her diagonally through the architecture, and the Spider Magnet reads walls as surfaces rather than obstacles.

None of that is decoration. InSuper Metroid (SNES, 1994) the pleasure was the moment your capability changed the map: a Grapple Beam turned a ceiling into a road, and a wall you’d walked past four hours ago became a door.Dread pushes that idea to its limit. Nearly every upgrade in the game is amovement upgrade — a new verb for crossing ground — and the ones that are weapons mostly function as keys. So the entire progression is one continuous statement: you have become better atgoing, and here’s a place that punishes anyone who can’t go well.

MercurySteam’s own history explains the second pillar. TheirSamus Returns (3DS, 2017) introduced the melee counter, andDread refines it into the game’s other axis. Enemies telegraph, you press the button on the beat, Samus knocks them into a free-aim slow-motion window, and you convert the parry into a kill. It’s a rhythm layer inside a movement game — the same instinct that runsSekiro, scaled down and applied to a corridor of frog-things.

The counter does something structural, though. It makesstanding your ground viable in a series whose combat has historically been a tax on exploration. Old Metroid enemies were furniture: you shot them, or you walked past them, and either way they were an interruption. Countering makes each one a tiny beat you can choose to play. That is the single largest change to Metroid’s minute-to-minute sinceSuper, and almost nobody talks about it because it’s not what the game was marketed on.

The E.M.M.I. zones, and why they work

They marketed it on the robots, and the robots are more interesting than their reputation.

E.M.M.I. are near-invulnerable machines that patrol sealed sections of ZDR. Enter their zone and the music drops, the map tints, and one of them starts hunting. Your beams do nothing. Caught, you get a single frame-perfect counter attempt with an extremely narrow window — succeed rarely, and die usually. The escape is the zone’s exit, and the exit is a door you have to reach while a machine that outruns you closes in.

Read as stealth, this is thin. There’s no meaningful hiding, the Phantom Cloak is limited, and the AI is more of a pursuit than a puzzle. Read aslevel design, it’s superb, and the reason is the same reason the counter works: it’s a tempo device.

An E.M.M.I. zone converts a map you know into a map you mustexecute. You’ve probably crossed that room before. You know where the exit is. What you don’t have is time to think, and so the zone tests the movement vocabulary the game just gave you — slide under, Flash Shift through, magnet up the wall — at speed, under pressure, with a fail state. It’s the Speed Booster puzzle with legs. And when you finally acquire the Omega Cannon and turn on the thing that’s been chasing you for forty minutes, the release is enormous precisely because the game spent forty minutes making you run.

The complaint that the zones break the pacing has it backwards.They are the pacing.Dread has no dialogue, few cutscenes, and no dramatic engine of any kind; the E.M.M.I. zones are the only device the game has for creating pressure and then removing it, and the reason the run feels like it has a shape is that the shape is made out of hunts and their endings.

The real ancestor is theMetroid II Metroid encounters — a hunt through a corridor with a countdown attached — given modern animation and an actual off switch.

Where it fights itself

Dread’s severity has costs, and the honest ones are worth naming.

The world is a machine rather than a place. ZDR is efficient, legible, beautifully signposted and slightly airless.Super Metroid’s Zebes had rooms that existed to be looked at, dead ends that were just dead ends, and a general willingness to waste your time in the interests of atmosphere. ZDR has almost no waste. Every corridor is a test, every room does a job, and the map is so well-designed that you’re rarely lost — which is a strange thing to complain about until you remember that being lost on Zebes was the entire experience. The signposting is so good it occasionally does your exploring for you.

Bosses run hot. The fights are excellent — sharp, readable, fast — and they arrive at a difficulty the rest of the game doesn’t prepare you for. Kraid is a wall. Escue and the Chozo Soldiers ask for a counter precision the corridors never demand. There’s a genuine gap between “very good at moving” and “very good at bosses”, andDread doesn’t build a bridge across it.

And the Aeion abilities — the resource-driven kit, Phantom Cloak and the rest — are the least-used thing in the game, for the same reason the prosthetics are the least-used thing inSekiro: the core loop is too complete to need them.

The verdict

Dread is the best-playing 2D game Nintendo has published in years and one of the most conservative, and those two facts are the same fact. It says a 2D Metroid is a movement game with a map on top, and then it executes that thesis with a precision the genre’s sprawling, generous descendants can’t match. It’s also the fastest-selling entry the series has ever had, which is a slightly funny outcome for a game whose main innovation is deciding what to leave out.

I’d hand it to anyone who thinks they like metroidvanias, because it will tell them whether what they actually like is the map or the running. Switch is the only place to play it, and it does not need a better machine.

I came to this from the wrong side. My 2D exploration-platformer of the eighties wasTurrican on the Amiga — Factor 5’s answer to this shape, all sprawl and firepower and no economy at all.Dread is the discipline that game never had, and playing them a lifetime apart makes the argument for restraint better than any review can.

For the descendants:Prince of Persia: The Lost Crown takesDread’s movement-first thesis and adds the systemic depthDread refuses, and it’s the better game for a lot of players.Animal Well goes the other way entirely and puts the whole progression in your head.

Spoilers below

Raven Beak is the argument’s conclusion, and he’s brutal. The final fight is a counter test — a long, escalating, four-phase demand that you read a Chozo warrior’s animation set with the precision the E.M.M.I. demanded and the corridors never did. It is the hardest thing in the game by a distance, and it lands because it is the only fight that asks for both halves of the toolkit at once: the movement to survive the arena, the counter to make progress in it.

The story around it is the partDread actually cares about, and it’s more pointed than the series usually manages. Samus’s Metroid DNA — theFusion inheritance, the thing that has been a plot device for two decades — becomes the resolution. She wins by being the monster the series has spent five games having her exterminate. That’s an ending with a real idea in it: the X parasites, the Chozo, the Federation and Samus herself are all the same story about a weapon that outlived the people who built it.

And then the game ends, cleanly, after roughly nine hours, with no post-game grind, no season, and nothing left to farm. Nineteen years of waiting for a thing that respects your evening. It closes the sentenceFusion started, and it closes it hard.

Most remakes are arguments. The remake ofResident Evil 4 takes its original’s central rule and overturns it.Final Fantasy VII Remake takes its original’s ending and makes a philosophical scene out of disagreeing with it. Motive’sDead Space, released in January 2023 on PS5, Xbox Series consoles and PC, does something rarer and much harder to talk about: it agrees. Fifteen years after EA Redwood Shores shipped the original in 2008, this is the same game, with the same layout, the same weapons, the same beats, executed by people who thought the 2008 design was already right and set out to build the version the hardware of the time wouldn’t allow.

That’s a strange product to review, because the temptation is to score it on novelty and it has almost none. The interesting reading is elsewhere.What does a remake look like when the designers’ only ambition is to remove the compromises? And how much of a horror game turns out to have been compromise?

Dismemberment was always the system

Dead Space’s reputation rests on one line of dialogue and one design rule, and the rule is the reason the game outlasted everything around it. Necromorphs don’t die from body damage. Empty a pistol into a torso and the thing keeps walking. You kill them by taking the limbs off, which means the plasma cutter’s rotating blade — horizontal, vertical, snap it round with a button — is an aiming problem rather than a damage problem. Every shot is a decision aboutwhich piece of a moving body you want to remove.

That does something no health bar can. It makes the enemy’s body legible. You look at a Slasher and you see arms, and arms are the problem, and the game has trained you to solve problems by cutting rather than by shooting. Panic inDead Space is the specific panic of firing four rounds into a chest because your hands forgot the rule.

The remake’s central technical addition serves exactly that rule and nothing else. Necromorphs are now built in layers — skin, muscle, bone — and shots strip them progressively, so a limb visibly degrades before it comes off. In a game about damage numbers this would be gore for its own sake. In a game where the enemy’s silhouette is your information, it’s afeedback improvement: you can now see how close a limb is to separating, and adjust mid-encounter. The 2008 game had a binary — attached or gone — because that’s what a 2008 console could stream. The remake has a gradient, and the gradient is the extra sentence the original wanted to say.

That’s the pattern for the whole project. Find the thing the design was reaching for, and give it the hardware it needed.

The ship as one continuous object

The USG Ishimura in 2008 was twelve chapters connected by loading screens dressed as tram rides. In 2023 it’s one place. You can walk from the bridge to the mining deck without a cut, the tram is a real vehicle in a real network, and — this is the part that matters — the ship is nowfully interconnected, so the level design can double back on itself the way a real derelict would.

Motive uses that in two ways. The obvious one is atmosphere: the Ishimura reads as a working vessel with a plan, which makes the corpses read as an event that happened to a place rather than a set of horror rooms in a row. The subtler one is the security clearance system, which replaces the original’s locked doors. Doors that refuse you at clearance one open at clearance two, so the ship gates you byrank rather than by scripted key, and the map fills in gradually as a single expanding space. It’s the metroidvania grammar applied to a survival horror ship, and it’s the single biggest structural improvement over the original.

Side missions follow from it. Isaac now has a handful of optional errands — chasing the Ishimura’s dead crew through their own logs — that send you back through territory you cleared hours ago. That trip is where the game earns the seamless ship, because returning to a “safe” corridor and finding it repopulated is a feeling the 2008 version’s architecture simply could not produce.

The Intensity Director sits on top, adjusting spawns, lighting and sound to what you’re doing. It’s less of a headline than the marketing wanted, and its real function is modest and correct: it stops the backtracking from being empty, and it keeps the ship from settling into a rhythm you can predict. The same idea, differently expressed, underwritesStill Wakes the Deep, where the horror is likewise a structure you have to keep re-crossing.

Isaac speaks, and the room changes

The one genuine deviation: Isaac Clarke has a voice, provided by Gunner Wright, who played him inDead Space 2 and3. In 2008 he was a mute — a helmet with hands, in the Gordon Freeman tradition.

This is the change most likely to annoy purists and it’s defensible on the game’s own terms. The 2008 Isaac was silent while every other character in the fiction spokeat him, issued orders, and treated him as an engineer to be dispatched. Giving him a voice converts him from an instrument into a person who is being used, which is the story the original was already telling and could only tell in the third person. The remake’s Isaac pushes back, occasionally, and every time he does the power dynamic of the Ishimura’s chain of command becomes visible.

The cost is real. Silence was doing work — the helmet is one of the great horror designs precisely because it never told you what was behind it. Trading that for characterisation is a legitimate trade with a legitimate loss, and it’s the only place in the remake where I’d say the original still wins.

Where it fights itself

The remake inherits the original’s third act, and the original’s third act is the weakest part of a very good game. Zero-G is now full flight, imported fromDead Space 2, and it’s an enormous improvement over the 2008 point-and-jump version — the asteroid-shooting sequence, which was reviled for fifteen years, is finally playable. It’s also the point where the game becomes a shooter with a level select, and the Necromorph rule that made the first six hours brilliant gets buried under set pieces.

Kinesis and Stasis are still underused. Kinesis lets you pick up a severed limb and throw it as a spear, and it’s the most inventive economy in the game: your ammunition is the enemy. Stasis slows anything to a crawl and turns a panic into a puzzle. Both are fully realised and both are optional, because the plasma cutter is so good that most players will finish the game having barely touched either. When a toolkit’s best tools are elective, the loop underneath is either magnificent or too dominant, and here it’s both.

And the upgrade tree keeps the original’s node system, which was fiddly then and is fiddly now — a grid of sockets that mostly amounts to spending currency on numbers. The remake adds an upgrade path for weapon behaviour rather than pure stats, which helps a little. It’s the one place Motive should have argued and didn’t.

The verdict

Dead Space (2023) is the most useful remake anyone has made, precisely because it’s the least interesting to argue about. It’s a demonstration that a well-designed game from 2008 needed nothing except the machine it deserved — that dismemberment, the Ishimura, and the plasma cutter’s rotating blade were finished work, and the loading screens were the flaw. Every change serves the original’s intent. Nothing is here to make a point.

The first six hours are as good as survival horror gets. The last three are a competent action game with a lot of set dressing, which is exactly what the last three hours of the 2008 version were, and I’d rather have the honest reproduction than a fabricated improvement. Play it on PS5 or PC; the ship is worth the frame rate.

If you want the opposite philosophy — a remake that disagrees with its source in public — readResident Evil 4 (2023). If you want the argument that survival horror’s soul lives in restraint rather than technology,Signalis makes it with a fraction of the budget.

Spoilers below

The remake’s smartest addition is the smallest one. In 2008, Nicole was a ghost you didn’t clock until the game told you, and the reveal at the end — that she’d been dead the whole time, that Isaac had been talking to a Marker-induced hallucination — landed as a twist because the fiction had withheld its cards.

Motive rebuilds it as something you can catch. The new-game-plus alternate ending, and the seeding of Isaac’s instability through the run, turn the reveal into a piece of evidence rather than a rug-pull. Isaac’s dialogue lets him respond to Nicole, and the responses go wrong in ways an attentive player registers long before the confirmation arrives. That’s the strongest justification for the voice: a silent Isaac cannot be seen losing his mind in real time, because losing your mind is a thing that happens in speech.

The Hive Mind is still a boss fight in a genre that shouldn’t have boss fights. It was true in 2008. It’s true now. Some things a restoration honestly cannot fix, and pretending otherwise would have made this a different, worse project.

The originalResident Evil 4 would not let you walk and shoot at the same time. Raise the gun and Leon planted his feet like a man in wet cement; the laser sight came up, the world narrowed, and the Ganados kept coming. That restriction was not a technical limit — the GameCube could have moved him — it was the design. Shinji Mikami’s 2005 game was built on a single, brutal trade:you may aim or you may leave, and you must choose now. Every encounter in that game is a clock made of that choice.

Capcom’s remake, released in March 2023 across PS4, PS5, Xbox Series consoles and PC, lets you move while aiming.

That’s the review, really. Everything else follows from it. Directed by Yasuhiro Anpo and Kazunori Kadoi under producer Yoshiaki Hirabayashi, and built in RE Engine after the studio’s runs atResident Evil 2 and3, the 2023RE4 is a genuinely superb action game that has a running disagreement with the thing it’s remaking. It knows what the original’s tension was made of. It takes it apart anyway, and then spends forty hours building a different tension in its place. Whether that trade lands is the only interesting question here, and it’s a closer call than the near-universal praise suggested.

What the stop-and-shoot rule was actually for

Give the 2005 design its due before touching it. Planting your feet did three jobs at once.

It madepositioning a decision made in advance. You had to pick your ground before the shooting started, because once you committed, you were furniture. That’s why the village opening works: the whole encounter is you reading a farmyard and choosing where to stand thirty seconds before you need to have chosen.

It madethe crowd frightening without needing the crowd to be fast. Ganados shamble. They shamble at a speed that is trivially outrun and lethal to a man standing still, and the entire threat curve ofRE4 is calibrated on that gap.

And it madethe shot expensive. Ammunition was scarce, so aiming cost you time; time cost you space; space was the only defence. Three resources, one input, all traded against each other. It’s one of the tightest economies action games have ever run.

Take the plant-and-shoot away and you have to replace all three, and Capcom knew it. The remake’s answer is the knife, and the answer is smart.

The knife as the new economy

In the original the knife was a tool: break crates, chip an enemy, cut yourself free. In the remake it’s the load-bearing defensive system, and it’s the best thing in the game. You can parry with it — including, remarkably, a chainsaw — you can stab a downed Ganado to stop them getting back up, you can slit a throat from behind, and you can cut yourself out of a grab.

And it wears out. Knife durability is the pivot the whole design turns on. Every parry, every finisher, every escape spends the blade, and the merchant charges to repair it. So the game reinstates the original’s core bargain in a new currency:defence is not free, and the resource it costs is one you have to buy back. You’re no longer trading time for space. You’re trading blade for safety, and a blade broken at the wrong moment leaves you doing the thing you spent the last hour avoiding, which is fighting a crowd with a handgun and no way out of a grab.

Around that, the tuning is meticulous. Leon moves while aiming, so the enemies got faster and more numerous, and they flank properly now. Parades of Ganados throw hatchets and sickles at range. The crowd pressure is genuinely higher than 2005’s because the game knows you can back out of it. The village fight is still the village fight — read the ground, pick your building, watch the ladders — and it’s harder, because standing still is now achoice rather than a consequence, and the game punishes the ones who make it lazily.

The dynamic difficulty system returns underneath, adjusting spawns and drops to how you’re doing, which is whyRE4 has always been the most forgiving hard game on the shelf and why almost nobody notices.

Ashley, and the argument Capcom won outright

The one change nobody sensibly disputes: Ashley Graham has no health bar. She has a downed state you can revive her from, no ammunition to manage, and a two-verb command set — follow or wait. In 2005 she had health, could be killed, and was the reason a generation of players remember an escort mission with a full-body shudder.

The remake’s version works because it identifies what the escort wasfor. Ashley’s job is to be a constraint on your movement and your attention — a thing that makes you look over your shoulder and route around a room differently. Her job was never to be a fail state you couldn’t control. Removing the health bar keeps the constraint and deletes the arbitrariness, and it is a masterclass in understanding your own game well enough to know which parts were load-bearing. Compare it with whatFinal Fantasy VII Remake does when it reaches the same fork: that game changes the thing to make a point about memory. This one changes the thing because the thing was bad.

The QTEs go too, mostly, and good riddance — the cutscene button prompts that killed you for blinking were 2005’s worst habit, and the remake keeps only the parry, which is a real mechanic rather than a reflex tax.

Where it fights itself

The remake is longer, denser, more generous and, in a very specific way, less frightening.

The originalRE4 had arhythm of pressure and relief that came from the movement rule. Between fights you were safe, because walking away was always available and always sufficient; inside fights you were trapped, absolutely, by your own trigger finger. The remake smears that. You can always move, always reposition, always slice your way out if the blade holds. That produces an unbroken, competent, slightly even hum of tension where the original had spikes and troughs. It’s a betteraction game and a slightly flatter horror one.

Then there’s the sheer amount of stuff. Side treasures to combine, spinels to trade, a shooting gallery, requests pinned to walls, weapon cases with attaché-case grids. The merchant’s economy is deep and enjoyable and it is also, hour by hour, the thing that keeps pulling your eye out of the world and into a spreadsheet. The original was a corridor with a shop in it. This is a shop with a corridor attached.

And a few of the strange edges got sanded. The remake reworks Salazar’s laser corridor into a stealth sequence, drops the U-3 fight, and generally trades 2005’s cheerful arbitrariness for coherence. Coherence is worth having. So was the arbitrariness, occasionally. The best moments in the original were the ones where a Spanish village stopped making sense.

The verdict

Capcom made the finest version of this game and, in doing so, made a case that Mikami’s constraint was a compromise rather than a thesis. I think that’s about sixty per cent right. The remake is better designed, better paced across its middle third, better at crowds, and vastly better at Ashley. The original is still the more frightening object, because being unable to move while a chainsaw approaches is a specific and irreplaceable feeling, and no amount of parry-timing gets you back there.

Take both. They’re arguing, and the argument is the interesting part. Available on PS5, Xbox Series and PC, withSeparate Ways — Ada’s parallel campaign, added in September 2023 — worth the money for the grappling hook alone, and the free Mercenaries mode from April 2023 sitting there as the purest distillation of the new combat system.

If you want to see a remake that takes the opposite position and restores rather than argues, go toDead Space (2023). If you want to see modern developers building a survival horror that keeps the old constraints on purpose,Crow Country is the one.

Spoilers below

The Krauser knife fight is where the remake’s whole design philosophy stands up and takes a bow. In 2005 it was a QTE — a sequence of button prompts flashed over a cutscene, memorised rather than played, and one of the most-hated three minutes in the game precisely because it asked for reflex rather than skill. The remake rebuilds it as a real knife duel: parries, reads, a proper contest fought with the mechanic the game spent thirty hours teaching you.

That’s the case for the defence in a single encounter. The remake takes 2005’s laziest moment and turns it into a test of the systems it built, which is exactly what a remake ought to do and almost none of them manage.

The counter-case is Saddler. The final fight is fine, competent, well-staged and completely conventional, and the original’s version was also fine, and neither is the reason anyone remembersResident Evil 4. Both games peak in a farmyard in the first hour, with a bell ringing and a crowd walking away, and no amount of engine work changes that. The village is the game. Everything after it is a very long, very good encore.

There is a moment, somewhere in the first ten hours ofSekiro: Shadows Die Twice, when the noise resolves into music. Up to that point you have been doing what every FromSoftware game sinceDemon’s Souls trained you to do: circle, wait, punish, roll away, drink, repeat. Sekiro tolerates none of it. The dodge is bad on purpose. The healing is scarce on purpose. Backing off gives your opponent the one thing the whole design refuses to hand you, which is time to recover. And then the penny drops, the sword comes in, you press L1 on the exact frame it lands, and the game answers with a metal chime that is unmistakably a downbeat.

That chime is the thesis. Released by FromSoftware in March 2019 under Activision, directed by Hidetaka Miyazaki,Sekiro is filed on shelves next toDark Souls andBloodborne, and it is a different genus entirely. The Souls games are about resource management under threat. Sekiro is about tempo. Once you hear it that way, everything from the posture bar to the resurrection mechanic to the game’s famously unbudging refusal of a difficulty slider snaps into a single coherent argument.

The posture bar is the whole game

Every FromSoftware combat system before this one ran on a health bar and a stamina bar, and your job was to spend the second to drain the first. Sekiro keeps health — vitality — and quietly demotes it. The bar that matters isposture, and posture is a balance meter. Land a hit, chip it. Deflect a hit, chip it harder. Let it fill and your opponent’s guard breaks, exposing them to a deathblow: one animated, unambiguous, fight-ending strike.

The clever part is what posture does when nothing is happening. Itrecovers, and it recovers faster the healthier the character is. That single rule is the engine of the entire game. It means backing off to breathe hands your progress back. It means the correct way to kill a boss is to stay inside their reach and keep the pressure on, which is precisely the thing a decade of Souls play taught you never to do. Sekiro spends its first act unteaching you, and the unteaching hurts, because the muscle memory it is fighting is your own.

So the posture bar converts defence into offence. A deflect — a block timed to the frame the blade arrives — costs you nothing, and costs them posture. Which means the optimal defensive action and the optimal aggressive action are the same button, pressed at the same moment. That is a rhythm system, dressed in Sengoku steel, and it explains why the game feels so bad until it suddenly feels superb. You are not learning tactics. You are learning a chart.

Why the deflect window works

Plenty of games have parries. Most of them make the parry a gamble: high risk, high reward, punished hard on a miss. Sekiro does something more generous and much more demanding. Mistime a deflect and you still block — you eat posture damage rather than a wound. The failure state of a perfect input is a mediocre input. That’s a design decision with enormous downstream consequences, because it means the game can hand you a boss who attacks in eight-hit chains and expect you to hold the line through all eight. You will not die from the third one. You will die from your posture cracking on the seventh.

Sitting on top of that are the perilous attacks, flagged with a red kanji, and they are the game’s genius stroke. Each one demands adifferent answer, and the answers are not interchangeable. A thrust wants the Mikiri Counter — step into the spear, not away from it. A sweep wants a jump. A grab wants your legs. The red kanji flashes with barely enough warning to react, which means you cannot read the symbol and then decide; you have to have already learned the animation that precedes it and be committed before the warning arrives. The symbol is a confirmation, not an instruction.

That is exactly how a rhythm game’s approach notes work. The note tells you what to hit and when, and by the time you consciously register it, your hands are already moving. Sekiro isGuitar Hero where the chart is a man with a naginata and the fail state is being taken apart at your own hearth.

Look for the real ancestor and you land onPunch-Out!!, where every opponent was a fixed loop of tells to be memorised and answered, and where “getting good” meant learning a script rather than raising a stat. I’d file the C64 and Amiga fighting games of my teens in the same family —The Way of the Exploding Fist,IK+ — where the whole contest lived in a single well-timed input and the loser was the one who twitched early. Sekiro is that lineage given twenty-five years of animation budget. Its cousins are elsewhere on this desk:Nine Sols takes the deflect and rebuilds it in two dimensions, and does it well enough to be worth the comparison rather than embarrassed by it.

The resurrection mechanic, and what it actually costs

Sekiro lets you die and get back up. Press the button, Wolf rises where he fell, and the fight continues from the enemy’s current state. On paper it’s a mercy. In practice it’s the most interesting piece of tuning in the game, because standing back up mid-fight puts you exactly where you least want to be — inside a boss’s active attack chain, at low health, with your posture recovery crippled. The revival hands you a chance and a worse position at the same time.

It also feeds the Dragonrot system, the game’s tax on repetition. Die often and NPCs across Ashina fall ill, sidequests stall, and your Unseen Aid — the chance of keeping your money on death — drops. Dragonrot is curable, and it is not really a punishment so much as a slow, visible pressure. The world coughs when you fail. It is one of the few times a FromSoftware game has made death mean something narratively without making it mean less mechanically.

The upgrade economy runs on the same honesty. Vitality and posture rise from Prayer Beads, and beads come in fours, from minibosses. Attack power rises from Battle Memories, and memories come from bosses. Healing capacity rises from Gourd Seeds. There is no build to hide behind. InElden Ring a wall can be walked around, out-levelled, or answered with a summon and a bleed build. Sekiro’s wall is a person, and the only thing that gets you past them is that you have learned the song. That is either the purest thing FromSoftware has ever made or the most obstinate, and the honest answer is that it’s both.

Where it fights itself

The prosthetic tools are the weakest system in the game, and it’s instructivewhy. Firecrackers, the Flame Vent, the Loaded Axe, the Umbrella — they’re inventive, they’re beautifully animated, and most players find two they like and never touch the rest. The reason is structural: the deflect loop is so complete, so self-sufficient, that the tools have nowhere to sit except as situational counters to specific enemy types. The game builds a second toolkit and then designs a combat system that doesn’t need it. The Loaded Umbrella against the Guardian Ape’s terror scream is genuinely essential; most of the rest is decoration on a machine that runs fine without it.

Stealth has a similar problem. Sekiro gives you a grappling hook, a crouch, ledge-hanging and one-hit backstab deathblows, and it’s a fine way to strip a boss arena of its rank-and-file before the real fight begins. It is also almost entirely optional and almost entirely absent from the encounters that define the game. The stealth exists to get you to the rhythm section.

The difficulty argument that swallowed the internet in 2019 looks different from here. FromSoftware shipped no difficulty options and took a proper kicking for it, and in October 2020 patched in Reflection of Strength for boss rematches and the Gauntlets of Strength — more ways to practise the chart rather than ways to lower it. Whatever you think of the accessibility question, the studio’s position was at least coherent with the design. You cannot ease a rhythm game without changing the song, and the song is the game.

The verdict

Sekiro is the most focused thing FromSoftware has built, and focus cuts both ways. It offers one answer to one question and asks you to get very good at it, which produces a stretch of about six hours in the middle where nothing works and you suspect the game is broken, followed by a click, followed by some of the finest one-on-one combat ever animated. Isshin, Genichiro, the Guardian Ape — these fights land because you arrive at them as a different player from the one who started, and you can feel the difference in your hands.

It doesn’t do the thing the Souls games do, where a hundred people play the same boss a hundred ways. Everyone beats Sekiro identically, because there is one way. Whether that’s purity or narrowness depends entirely on how much you want a game to hold an opinion. This one holds a very strong one, and it’s right.

Play it on PC or PlayStation; it runs fine on both and the frame timing is the whole product, so give it stable hardware. If it takes and you want the same grammar in a different key,Nine Sols is the sharpest descendant going, andArmored Core VI is FromSoftware doing the same trick with a stagger bar and a rocket launcher.

Spoilers below

The design argument closes at Isshin, the Sword Saint. Three phases, no gimmick, no environmental trick, no summon — just the game asking whether you learned it. The first phase is Genichiro again, which is a joke at your expense and a genuinely kind one: the man who annihilated you in hour three is now the warm-up. Isshin’s second phase adds the spear and the thrusts, and every thrust is a Mikiri check. The third adds lightning, and the lightning is the only moment in the whole game where a piece of the toolkit outside the deflect loop becomes mandatory — the Lightning Reversal, learned from a fisherman, thirty hours earlier, in a place you probably haven’t thought about since.

That’s the argument landing. Sekiro’s final exam tests one skill, and the skill is listening. Everything else — the tools, the stealth, the Dragonrot, the Sculptor’s grief — is set dressing on a metronome. The Shura ending, where Wolf turns on his own oath, is thematically the right shadow to hang over a game about obedience to a rhythm. You do what you are told, on the beat, for forty hours. Then it asks whether you can stop.

There’s a moment inUmurangi Generation where you’re on a rooftop in Tauranga composing a shot of some graffiti, and you notice — properly notice, for the first time — the thing standing on the horizon behind it. The game hasn’t cut to it. No camera swing, no stinger, no character pointing. It’s been there since you loaded the level. You just hadn’t looked, because you were doing your job, which was photographing graffiti for money.

That’s the whole design, delivered in one gesture.Umurangi Generation is a game about the difference between looking and seeing, and it teaches that difference by paying you to look.

The setup

Released in May 2020 on PC by ORIGAME DIGITAL — essentially Naphtali Faulkner, a Māori developer working out of Australia — with a Special Edition arriving on Switch in 2021 via Playism and consoles following in 2022. The title is te reo Māori:umu rangi, red sky. It won the Seumas McNally Grand Prize at the Independent Games Festival in 2021, which is the closest the indie scene has to a Best Picture, and it beat a field with far more money in it.

You’re a courier with a camera. Each level drops you in a location in a near-future Tauranga, hands you a bounty list — photograph this, photograph that, photograph three of these in one frame — and a time limit. Complete the list, earn gear: new lenses, new film stock, filters, a tripod. Then the next level, which is worse.

Why the bounty list is the mechanic

The obvious ancestor isPokémon Snap, and it’s a real lineage: a game where the verb is framing and the scoring is composition. ButSnap is on rails and its subjects perform for you. Faulkner took the framing verb and put it in a space you walk, climb and clamber through, which changes what a photograph is. InSnap you receive a subject. Here you go and find one, and finding is the gameplay.

Now the trick. The bounty list is a checklist of banal nouns. Bins. A skateboard. Someone’s mate. A pigeon. It is deliberately, aggressively mundane, and it directs your attention like a lead in the nose. You are scanning the level for a bin.

Meanwhile the level is telling you a story. There are UN soldiers in the street. There are refugee tents in the car park. There’s a mural somebody painted about what happened to their neighbourhood. There’s the thing on the horizon. None of this is on your list, and none of this is required, and the game will never acknowledge that you saw it.

So the design does something almost no political game manages. It doesn’t lecture you. It gives you an errand, surrounds the errand with a catastrophe, and lets you be the person who chose what mattered. If you photographed only the bins, that’s information about you. The critique isn’t in the text. It’s in what your own attention did in a room.

That’s a much older trick than it looks, and it’s an environmental-storytelling one. The real ancestor is the wayHypnospace Outlaw makes you an enforcer clicking through webpages for copyright violations while somebody’s life falls apart in the sidebar. Both games weaponise a job description. You’re compliant, you’re being paid, and the compliance is what stops you looking up.

The gear is the point of view

Photography games usually treat lenses as stats.Umurangi treats them as positions. A long lens compresses a scene and flattens distance, which makes a crowd look like a mass. A wide lens exaggerates space, which makes a soldier standing over a civilian look like architecture. Faulkner clearly knows this, and the levels are built so that the same subject reads differently depending on the glass you brought.

There’s also a photo mode that’s an actual photo mode: exposure, colour grade, depth of field. And critically, the game lets you take pictures that have nothing to do with the bounty. You can spend the entire timer photographing a wall. The timer is generous enough that this is viable and tight enough that it costs you.

This is the freedom that makes the politics land rather than nag. A game that forced you to photograph the atrocity would be a game telling you the atrocity is important. A game that pays you for bins while the atrocity is available in the background is a game asking what you’d do with a camera and a wage.

The soundtrack, and the anger

ThorHighHeels did the music, and it’s a genuinely great record — jazzy, warm, loose — and it does the same work the bounty list does. It’s too pleasant for what you’re seeing. It’s the sound of a Tuesday. The dissonance between the mood of the audio and the content of the frame is where a lot of the game’s discomfort actually lives.

Faulkner has been direct in interviews about the origins: the 2019–20 Australian bushfires, the spectacle of institutions responding to a catastrophe with press conferences, and a specifically colonial reading of who gets protected when the sky turns red. The game inherits that anger without inheriting a thesis statement. Nobody in it makes a speech. The UN presence just gets more numerous, level over level, and the tents get more numerous, and eventually you’re photographing something that used to be a town.

The frame as a lie

One more thing this game knows that most photography games don’t: a photograph is an edit.

Every shot you take excludes almost everything. Step left and the soldier leaves the frame. Crouch and the tents disappear behind a wall. Zoom and the context evaporates.Umurangi never says any of this out loud, and it doesn’t have to, because it made you do it several hundred times. You have personally cropped a crisis out of a picture in order to get a clean shot of a bin, and you did it for a small amount of money, quite quickly, without thinking about it.

That’s the argument. Photography chooses, and choosing is a political act performed by somebody’s hands, and in this case the hands were yours. Games have spent decades trying to make the player complicit through plot twists. This one does it with a viewfinder and a shopping list.

Where it strains

It’s rough. The movement is loose, the clambering is inelegant, the collision occasionally embarrassing, and there’s a level or two where finding the last bounty item is genuinely tedious rather than observant. This was made by essentially one person and it plays like it in the seams.

It’s also short — two to three hours for the base game, a bit more with theMacro DLC, which is the best content in the package and considerably angrier than the main campaign. Short is the right shape. It’s still worth knowing.

And the bounty design occasionally fights the looking. When a list item is fiddly — get four of these in one frame from a spot that barely exists — you stop being a photographer and start being a checklist operator, which is the exact mental state the game is critiquing, achieved by accident rather than design.

The verdict

Umurangi Generation is the most efficient political game I know, and the efficiency is the achievement. Three hours. One verb. No dialogue trees, no morality meter, no branch where you choose to Be Good. It just hands you a camera, gives you a reason to point it at something trivial, and puts the end of the world in the depth of field.

The medium keeps trying to do politics through writing — a character explaining the system, a choice menu about the system — and keeps producing homework. Faulkner did it through attention, which is the one resource games actually control. He made noticing optional and then measured nothing, and that’s why it works. The game never tells you that you missed it. You just find out later that it was there.

Play it on PC if you can; the mouse is the camera and the camera is the game. The Switch version is a competent port and the right size for a couch, though the photo-mode fiddling is happier with a pointer.

Where next:Hypnospace Outlaw for the other great game about doing a small job inside a large disaster.NORCO for a place that has already had its red sky and learned to live under it.

Spoilers below

The escalation is the structure, and it’s brutal once you see it laid out. Early levels are a skate park and a hangout — you photograph mates, you photograph a crew. By the middle, the same locations have soldiers in them. By the end you are photographing the aftermath of something that killed people you’d previously been asked to photograph having a nice time, and the bounty list is still asking for bins.

The list never changes tone. That’s the knife. The game could have had your employer stop, or apologise, or pivot to documenting the crisis, and instead the errands continue at the exact register they started at, because the institution issuing them does not have a mechanism for noticing. The horizon fills up and the paperwork stays the same shape.

And the final level’s use of the camera — where the only thing left to photograph is what happened — works because you’ve spent three hours with the shutter making a small pleasant sound. Two hundred photographs of bins have taught your hand a reflex, and the game finishes by pointing that reflex at the thing it was always in the way of.

About four hours intoParadise Killer I worked out what it had done to me and had to put the pad down for a minute.

I was standing on a beach on a dead island, holding evidence that pointed at somebody. Not conclusively. It pointed. And the game’s interface was telling me, as it had been telling me since minute twenty, that I could go to trial with it right now. No gate. No “you need more clues before you can proceed”. The Judge would convene, I would present exactly what I had, and something would happen.

Detective games do not do this. Detective games check your work.

What it is

Paradise Killer came out in September 2020 for PC and Switch, from the small British studio Kaizen Game Works, published by Fellow Traveller, with PlayStation and Xbox versions following in 2022. It’s a first-person open-world investigation set on Island Sequence 24, the twenty-fourth in a series of artificial islands built by a cult to resurrect dead gods, each one eventually corrupted and abandoned so the next can be built.

The entire ruling Council has been murdered on the eve of the island’s retirement. You are Lady Love Dies, an “investigation freak” who has been in exile for three million days and gets recalled to solve it. You have a computer companion called Starlight, a currency of blood crystals, and total freedom of movement across a vertical vaporwave ruin you’re expected to climb by finding a fast-travel network and a set of movement upgrades.

The soundtrack, by Barry “Epoch” Topping, is city-pop and lounge and it is the best argument the game makes for itself in the first ten minutes.

Why removing the right answer works

Every detective game before this one has a correct solution and a verification step.Ace Attorney will not let you present the wrong evidence — you get a penalty and a retry.Obra Dinn confirms in threes.Golden Idol tells you flatly that your sentence is wrong. All three are excellent, and all three share an assumption: the game knows, and your job is to converge on what the game knows.

Kaizen removed the verification. There is a truth — the game has a real answer to what happened — and the trial does not require you to have found it. You accuse who you accuse, with what you’ve got, and the trial resolves accordingly. People are sentenced. Possibly the wrong people.

The effect of this is not chaos. The effect isresponsibility, and it changes what investigating feels like at a physiological level. When a game verifies you, evidence is a key: does it fit, yes or no. When a game won’t verify you, evidence becomes an argument you are choosing to make about a person, and you feel the weight of the choice while you’re making it. I found myself doing something I have never done in a detective game: going back out for corroboration I didn’t need to progress, because I wasn’t sure enough to say it out loud.

That’s the design win. The freedom to be wrong converts a puzzle into a judgement, and judgement is the thing the fiction of detective work is actually about.

The island as an evidence board

The other half of the design is spatial, and it’s the half that gets undersold.

This is an open world with no combat, no enemies and no icons dumped on a map. It’s a large vertical space with clues embedded in geometry — on rooftops, under walkways, at the end of climbs the game never signposts. Movement upgrades are purchased from a vendor with blood crystals you find by exploring. So the loop is: explore to afford mobility, use mobility to explore.

That’s a Metroid economy wearing a detective’s coat, and it’s why the island reads as a crime scene rather than a hub. The knowledge you accumulate isn’t only propositional — “the Marshal was seen here at this hour” — it’s geographic. You learn that two locations are closer than the suspects claimed because you climbed between them. Testimony collides with architecture. When a character’s alibi depends on a distance, you have legs and you can check.

Compare whatAnimal Well does with a world that has to be understood before it can be traversed. Same instinct, different genre coat: the map is the puzzle and the puzzle is the map.

Where it fights itself

The dialogue is a lot. Kaizen have committed hard to a register — cult jargon, proper nouns with capital letters, characters named Doctor Doom Jazz and Crimson Acid — and the game’s density of invented vocabulary in the first hour is a real barrier. Some players bounce off before the systems get a chance to show what they’re for. That’s a legitimate cost of the aesthetic and worth naming rather than excusing. The world-building is coherent, and coherent is not the same as welcoming.

The interrogations are also structurally repetitive. You visit a suspect, you fan out your evidence, you tick topics off. There’s no pressure mechanic, no lie detection, no risk in the room. Given how bold the trial is, the conversations leading to it are conventional in a way that mildly undercuts the whole.

And the trial itself is more presentation than combat. You lay out your case and the Judge processes it. It is dramatically flat compared with what precedes it — though I’ve come around on this. A theatrical trial would have suggested the game was scoring you, and the game’s entire thesis is that it isn’t.

The thing it understands about detective fiction

Worth putting plainly, because it’s the insight the rest of the genre keeps missing.

A detective story has two engines. One is the puzzle — the impossible room, the alibi that doesn’t hold, the timetable. The other is the detective’s authority: somebody decides what happened, and their deciding is what converts a mess of facts into a public truth. Christie runs on the first. Chandler runs on the second. Games have, almost without exception, only ever built the first, because the first is a lock and games know how to make locks.

Kaizen built the second. The puzzle inParadise Killer is honestly middling — the clues are findable, the chains aren’t fiendish, and a careful player will get there. What’s exceptional is that the game models theact of concluding as a thing with consequences that belong to you. That’s why an average mystery produces an above-average detective game. The mystery was never the interesting part; the deciding was.

You can watch other designs circle this.Immortality hands you footage and no verification and gets somewhere adjacent by making interpretation the mechanic.Paradise Killer is the version where interpretation has a defendant.

The verdict

Paradise Killer is a tiny-team game with an idea that a hundred-person studio would have focus-tested into the ground. It found the load-bearing convention of its genre — the correct answer — pulled it out, and demonstrated that the building stands up better without it.

The island helps. Vaporwave is a style that has aged into wallpaper over the last decade, and this is one of the few games that had a reason for it: an artificial paradise built by a cult, dressed in the aesthetic of a future that never happened, on its twenty-fourth attempt. The pastel decay is an argument about the setting rather than a mood board. Ruins with palm trees and a synth bass are exactly what a failed utopia would leave behind.

Twelve to fifteen hours if you’re thorough, and thoroughness is the mode it wants. It’s on PC, Switch, PlayStation and Xbox. PC with a mouse suits the reading; the Switch version is the one I’d hand to somebody who wants to sit with it, and the soundtrack is worth a decent pair of headphones either way.

Where next: if you want the same evidence-assembly rigour with a stricter marker,Return of the Obra Dinn remains the high-water mark. If it’s the interrogation-as-character-study you want,Disco Elysium does what Paradise Killer’s conversations gesture at and don’t reach.

Spoilers below

The reveal that Lady Love Dies was exiled for a reason — and what that reason turns out to be — reframes the freedom to accuse anyone into something considerably darker. The game hands you unlimited prosecutorial power and then discloses that your character has previously used judgement badly enough to be removed from the world for three million days. You are the least qualified person on the island to be doing this, and the Council appointed you anyway, because the Council needed somebody who would deliver a verdict rather than the truth.

Which is what makes the ending options land. You can convict the wrong person knowingly. Not by accident, not by failing a check — you can look at the real answer, decide the island is better served by a different one, and file it. The game permits it and then makes you watch the sentence carried out. There’s no punishment screen. There’s no correction. The island simply continues on the version of events you signed.

Doctor Doom Jazz, Crimson Acid, the Marshal, every suspect I spent hours picking apart — the game’s real position is that the Syndicate was always going to build Island Sequence 25 regardless of who I named, and my investigation was a procedural formality performed to make a machine feel legitimate. That’s a hell of an argument to smuggle in under the vaporwave.

The detective genre in games has spent thirty years trying to make deduction happen and mostly producing its opposite. You know the pattern: you walk into a room, press the button on every glowing object, and once the counter reads 6/6 the detective announces the solution he worked out without consulting you. The game calls this an investigation. What it actually is is a search-and-collect with a lecture at the end.

The Case of the Golden Idol, released in October 2022 by the small Latvian studio Color Gray Games and published by Playstack, does the obvious thing that almost nobody does. It gives you the evidence and then makes you say what it means. If you’re wrong, it says no. It does not say why.

That “it does not say why” is the entire product.

The mechanism

Each of the eleven scenes is a single tableau: a frozen moment, hand-drawn in a style somewhere between Hogarth and a bad dream, populated by grotesques mid-crime. Somebody is falling off a cliff. Somebody is being poisoned. You click around the scene — pockets, letters, ledgers, signage, faces — and every clickable thing yieldswords. Names. Occupations. Verbs.

The words go into a bank. Then you open the thinking panel, which is a page of sentences with holes in them, and you drag words into holes until the sentences describe what happened. Who is who. Who did what to whom, and with what, and why.

That’s it. There’s no dialogue. There’s no interrogation, no timeline scrubber, no notebook that fills itself in. Two verbs: look, and assert.

Why the word bank is smarter than a dialogue tree

Here’s the design problem every detective game hits. Deduction is internal. It happens in a head. To make it a mechanic, you have to externalise it, and the moment you externalise it you risk turning “I worked it out” into “I picked the right option from three”.

The word bank solves this by making the answer spacecombinatorial and hostile. When a puzzle has forty available nouns and eleven slots, brute force isn’t a strategy — it’s a punishment. You can’t guess your way through, because the possibility space is too wide to walk and too narrow to fluke. So you’re pushed back into the only remaining approach: actually thinking about it.

And the game refuses to grade partially in a way that would let you triangulate. This is where it separates from its most obvious relative.Return of the Obra Dinn confirms your fates in batches of three, which is a genuinely brilliant compromise — it stops the game being unwinnable while making you commit to trios. It also means a canny player can farm it: lock two you’re sure of, cycle the third.Golden Idol declines the compromise. Submit an imperfect answer and you learn that it’s imperfect, and you go back to the tableau with your ego intact and your theory in pieces.

The result is that the moment of solving is undiluted. Nothing helped you. The game withheld everything except the facts, and the facts were sufficient, and you found them sufficient. I can’t think of a cleaner delivery of that feeling in the medium.

The other trick: the story is in the ledger

The eleven cases run across decades, and the plot — a cursed golden idol, an inheritance, a family, a great deal of murder — is never narrated to you. It’s assembled from the same nouns you’re using as puzzle pieces. You learn the dynasty’s shape because you keep filling in surnames. You work out the political situation because a scene requires you to identify who signed a document.

This is a genuinely rare thing: exposition that costs the player effort and therefore sticks. Nobody remembers a cutscene. Everybody remembers a name they had to earn. It’s the same economicsTunic runs when it makes the manual pages both the lore and the solution, and it’s why both games feel dense at a fraction of the word count of a proper RPG.

The art carries more of this than it gets credit for. The figures are ugly on purpose — pop-eyed, jowly, caught mid-gesture — and the ugliness is functional, because you need to distinguish nine strangers at a glance across ten scenes with no name tags. A realistic style would have made them a soup. Caricature is a legibility tool that happens to also be a tone.

Where it fights itself

Two honest complaints.

The scenes are static, which means the tableau has to carry both the puzzle and the drama, and occasionally the drama loses. A frozen frame is a fantastic puzzle substrate and a limited storytelling one, and a couple of the mid-game cases feel like admin — identify eight people at a party — rather than a crime you care about.

And the difficulty is uneven in the way hand-built puzzle games always are. Most of the eleven land beautifully. One or two hinge on a single obscure noun in a corner, and if you don’t click that corner you’re not stuck on logic, you’re stuck on pixel hunting, which is a different and worse kind of stuck. The game has a hint system for exactly this, and using it feels like a small defeat, which is arguably correct and definitely annoying.

The 2023 DLC chapters — The Spider of Lanka and The Lemurian Vampire — are tighter than the base game on both counts, which is a good sign about what the studio learned. The 2024 sequel,The Rise of the Golden Idol, moves the whole apparatus forward a couple of centuries and adds quality-of-life the original lacked.

The bit about being wrong

I want to dwell on failure, because it’s the least discussed part of this design and the most radical.

Modern games treat a wrong answer as a UX problem. Something must happen: a hint surfaces, a difficulty slider quietly nudges, an NPC wanders over to helpfully observe that the lever looks operable. The industry spent twenty years engineering frustration out, and in the process engineered out the state that precedes insight. You can’t have the click if nothing was stuck.

Golden Idol lets you be stuck. Properly, unproductively, for a quarter of an hour, staring at a picture of a man in a wig. And the reason this is tolerable rather than infuriating is a quiet piece of craft: the scene is always complete. Everything you need is on screen. There’s no second location, no locked area, no character who’ll say the missing thing on Tuesday. So when you’re stuck, you know with certainty that the failure is comprehension. That certainty is what makes persistence rational.

This is the oldest lesson in the medium and it keeps getting mislaid. The C64 adventures I grew up on were frequently stuck-forever affairs, and the good ones differed from the bad ones on precisely this axis: whether the puzzle was closed. A closed puzzle you can’t solve is a challenge. An open one is a guess. Color Gray have simply remembered which is which.

The verdict

Golden Idol is the rare game that respects you by ignoring you. It won’t encourage you. It won’t nudge. It has no interest in your session length or your completion funnel. It puts a horrible little painting in front of you and waits.

The genre lesson underneath it is worth naming: detective games have been adding features — timelines, reconstructions, deduction boards with animated string — when the missing ingredient was always subtraction. Take away the confirmation and the thinking arrives on its own. Every mechanic Color Graydidn’t build is why the one they did build works.

It’s on PC, Switch, PlayStation, Xbox and phones. The phone version is better than it has any right to be — the whole game is clicking and dragging, and a tableau sits fine on a tablet. Play it in single-case sittings with a real pen if you’re that way inclined. Most people won’t be. Most people will find they need to be by case seven.

Where next:Obra Dinn is the sibling and the better game overall, though not the purer one. If you want deduction with an actual world to walk around in, and a game that will happily let you be catastrophically wrong,Paradise Killer is the other end of the same argument.

Spoilers below

The idol itself is the best-kept structural joke in the game. For eleven scenes you’re doing forensic work on a series of murders, and the object motivating all of them has a power that is never explained by any mechanism and never needs to be, because the game has correctly identified that its supernatural MacGuffin is doing zero puzzle work. The idol is a reason for people to be greedy. Greed is legible. Curses are furniture.

The dynasty structure — the way the same family line keeps regenerating the same crime across generations — pays off because you built the family tree yourself, one dragged surname at a time. When the last case asks you to name a relationship you established four scenes ago, it’s checking whether you were investigating or just solving. Those turn out to be different activities, and it’s the only game I know that can tell the difference.

The pitch forDredge sells it short in both directions. “Lovecraftian fishing game” makes it sound like a novelty — a cosy loop with tentacles glued on for the trailer. What Black Salt Games actually shipped in March 2023 is a compact resource-management design where the horror isn’t a layer on top of the fishing. The horror is what the fishing is made of.

Four people made it, out of New Zealand, published by Team17. It came out on PC, Switch, PlayStation and Xbox on the same day, which for a team that size is its own small achievement. It runs about twelve hours if you’re thorough. It has no combat. It has one of the tightest pressure systems of its year.

The loop, plainly

You have a boat. You sail to a shoal, play a short timing minigame, and a fish comes aboard. The fish occupies a shape in your hold — a grid, Tetris-style. You sail to a dock, sell the fish, buy an upgrade: bigger hold, better rod, a dredge for pulling scrap off the seabed, an engine. You take on requests from the people in the archipelago’s five regions, and the requests pull you into the story.

Read as a list, that’s a chore simulator. Played, it’s a vice. The reason is that every one of those verbs is competing for the same two scarce things: space and daylight.

Why the grid is the whole design

The hold is a grid, and everything lives in it. Fish take up squares. So do your rods. So does the dredge. So does the crab pot, the research parts, the cargo you promised somebody. This single decision does about four jobs at once.

It makes gear a cost. In most games, equipping a better rod is a strict upgrade — you press a button, the number goes up. Here the better rod is a bigger rectangle, and a bigger rectangle is fish you can’t carry. You are constantly asked whether capability is worth capacity, and the answer changes by the trip.

It makes fish individual. A cod is a shape. An aberration — a mutated fish, the kind you pull up at night — is an awkward shape, and it’s worth more, and it sometimes rots and infects its neighbours. So the hold becomes a small hostile puzzle you are packing under time pressure, and the horror gets to live in your inventory screen rather than in a cutscene.

And it makes the return trip a decision. Full hold, three hours of light left, a good shoal one leg further out. Do you push?

The real ancestor of this is the attaché case inResident Evil 4 — the 2005 one — which took the most boring screen in games and made it a place you’d voluntarily spend time. Capcom understood that if arrangement is a skill, storage becomes content.Dredge takes that insight and pushes it further by making the case the scoreboard. The RE4 case rewarded neatness; theDredge hold decides what your day was worth. If you want the full argument about how the 2023 remake handled that legacy, that’sits own piece.

The deeper ancestor isSunless Sea, which established the modern template: a small boat, a dark map, a hunger meter and a fatal curiosity about the next island.Dredge is the arcade edit of that game. It cuts the prose, keeps the dread, and replaces the slow-burn attrition with a clock.

The clock is the horror

Here is the mechanism, and it’s elegant enough to admire in isolation.

Sail at night and a panic meter fills. As it fills, things appear: rocks that weren’t in that water at noon, shapes at the edge of the light, something with a wake. Push far enough and you take damage from things the game has never formally introduced to you.

What makes this work is that the game never forbids it. Night sailing is permitted, and it is profitable — aberrations are night fish, and aberrations pay. The horror is therefore always somethingyou elected to do for money. That’s a substantially different feeling from a monster that arrives on a schedule. You are not being ambushed; you are being tempted, and the game is politely keeping a tally.

This is the same trickPacific Drive runs with the storm timer: the danger is a resource you spend against a reward, so every scary moment is retroactively your own fault. Fault is a much stickier emotion than fright.

The panic meter also solves the cosy-game problem. A cosy loop wants you to settle in. A horror game wants you unsettled.Dredge resolves the contradiction by putting them on the same axis — daylight is cosy, and darkness is the same activity with the safety off — so you get the calm and the terror out of one set of rules, and the transition is a slider rather than a door.

The small cruelty of the timing minigame

One more piece deserves credit, because it’s the bit reviewers skip. Catching a fish is a timing test: a marker travels, you stop it in a zone, repeat a few times. It’s slight. It is also deliberatelyfast, and it takes real seconds, and those seconds are daylight.

That’s the trick. In a game where light is the scarce resource, a minigame is no longer a diversion — it’s a meter running. Fumble the rhythm on a big fish and you’ve spent twenty minutes of in-world afternoon on a failure. So the minigame being trivial is the design working: it’s not there to be hard, it’s there to convert your attention into time, so that the sun going down is something you did rather than something the clock did.

Most games would have made the fishing test harder and the fish more valuable. Black Salt made it cheap and made it cost, which is why an eight-second interaction you’ll perform hundreds of times never quite becomes furniture.

Where it strains

Traversal. The archipelago is generous at first and repetitive by hour eight, because sailing is a constant-speed activity with nothing to do during it, and once you know the map you’re mostly holding a stick forward. Engine upgrades help. They don’t fix it. The Iron Rig expansion in 2024 loaded the mid-game with more to do, and The Pale Reach in late 2023 added an ice region with genuinely new fishing rules, but the base game’s middle hour has a slack patch and it’s honest to say so.

The story is thinner than the atmosphere. The Collector wants things; you fetch them; the fetching is meaningful and the character isn’t. This is a game with an extraordinary sense of place and a serviceable plot, and if you arrive expecting the writing to match the systems you’ll be a little cold on it. CompareSignalis, which inverts the ratio.

The verdict

Dredge is a design worth studying and a game worth finishing, and those are different compliments that it happens to earn together. Four people identified that a fishing game and a survival-horror game have the same skeleton — go out, gather, weigh the risk of one more, come home — and built one object that is both. The grid hold and the panic clock are the entire achievement. Everything else is decoration on a very good machine.

There’s a lesson in here for bigger studios that won’t take it.Dredge has one idea about space and one idea about time, and it spends its entire runtime compounding them against each other. No skill tree padding the middle, no crafting web, no map littered with icons to justify the map’s existence. Twelve hours, two mechanics, and the fact that I can still describe the whole thing in a paragraph is a feature of the design rather than a limit of it.

It’s on everything. Switch handheld suits it — this is a game for a chair — but the timing minigame is marginally kinder with a mouse or a decent stick, and the PC version is the one to pick if you’re chasing the expansions.

Where next: for the survival loop with better prose, takeMouthwashing; for the fuller inventory-as-anxiety experience, go back toCrow Country and watch the same instincts get applied to a mansion instead of an ocean.

Spoilers below

The Collector’s endgame — assembling the pieces of the artefact, and the choice the game offers once you have them — is whereDredge declines to be as smart as its systems. The dual ending is a lever pull. Having spent twelve hours teaching you that decisions are shaped like rectangles and priced in daylight, it resolves on a binary menu, which is the one moment where the fiction and the mechanics stop talking to each other.

The better ending — the one where you understand what the fisherman has been doing and what the pieces are for — lands emotionally because the aberrations have been telling you the whole time. The mutated fish aren’t set dressing. They’re the same process that’s happening to you, applied to something with fins, and the game has been putting the evidence in your hold and charging you money for it since hour one. You sold the symptom. That’s the joke, and it’s a good one.

What stays with me is that the horror had a price list. Every dreadful thing in this game arrived with a market value attached, and I kept taking the trade.

Norco is a real place. It sits in St. Charles Parish, Louisiana, about twenty miles up the Mississippi from New Orleans, and its name is an acronym: the New Orleans Refining Company, which built the town around the plant in the 1910s. The refinery is still there. Shell owns it. The town is named after the thing that poisons it, which is the sort of detail a fiction writer would be told to cut for being too on the nose, and it is simply the address.

That’s the first thing to understand aboutNORCO, the 2022 point-and-click adventure by Geography of Robots, published by Raw Fury. It didn’t invent its setting. It reported one. The developer grew up down there, and the game carries the specificity of somebody describing a place they can’t stop describing — petrochemical flare stacks, drainage canals, the particular light of a swamp that has industry sitting in the middle of it.

The shape of the thing

You play Kay, returning home after her mother Catherine has died of cancer. Her brother Blake is missing. That’s the engine: find the brother, settle the estate, leave. Nobody leaves.

Mechanically this is an adventure game of a very old school — pixel art, cursor, inventory, conversation trees, the whole 1990s apparatus, rendered with the muddy colour palette of a machine that had a limited number of colours and made a personality out of it. Kay is accompanied by Million, an android her mother owned, who narrates and comments and is funnier than the situation deserves. There are minigames. There is a mind map.

The mind map is the piece worth stopping on, because it’s the design decision that makes the game work.

Why the mind map works

Most adventure games track state in an inventory and a journal.NORCO tracks it in a diagram of Kay’s head: characters, places, ideas, connected by lines you can click to have Kay tell you what she thinks about the connection. It looks like a convenience feature. It is doing something much more specific.

An inventory tells you what you’re carrying. A mind map tells you what you’rethinking about, which in a game where the actual puzzle is “why is my family like this” is the only inventory that matters. Consulting it isn’t a lookup; it’s Kay worrying at something. The design is telling you that the obstacles here are not locked doors — they’re the things she hasn’t understood yet about her mother.

The real ancestor of this isDisco Elysium’s Thought Cabinet, which turned ideas into equipment, but the lineage runs further back than that. Anyone who played adventure games in the Amiga years remembers the topic-list dialogue system, where you’d hoover up keywords and try them on every NPC like keys on a ring.NORCO looked at that mechanism — the game’s memory of what you’d learned — and asked what happens if you make it the interface rather than the plumbing. It’s the same moveTunic makes with its manual: promote the paratext to the text.

The other thing the mind map buys is pacing control without gating. The game rarely stops you. It lets you carry confusion around, and it gives you a place to put the confusion down, which is why a story this dense doesn’t feel like homework.

The refinery is the antagonist

There’s no villain here in the sense a game usually means it. There’s a company, a security apparatus, some men with a lot of money and a religious streak, and an economy that has already decided what the town is for. The plot involves all of them. The pressure comes from something more diffuse: a place where the largest employer is also the reason the air tastes like that, and where leaving is expensive and staying is expensive and both bills arrive.

NORCO is very good at the texture of this. The Shell plant looms over dialogue the way weather does. Characters talk about work, and about who’s sick, and the two conversations are the same conversation. The game never delivers a thesis paragraph about extraction; it just keeps showing you the drainage ditch.

There’s a formal trick underneath it.NORCO keeps handing narration duties around — Kay, Million, Catherine’s recollections, the occasional block of text that belongs to nobody in particular — and the effect is that the town accumulates description faster than any single character could supply. Adventure games usually fix the camera to one consciousness because the cursor implies a hand. This one lets the perspective drift, and the drift is the point: a place gets described by everyone who’s stuck in it.

This is where the Southern Gothic label earns itself. The genre’s actual content — Faulkner, O’Connor, the decayed grandeur and the inherited guilt — is about a place where the past won’t decompose.NORCO relocates that to a landscape where the past is literally in the groundwater, and lets the mode do the work. There are visions. There are prophets. There’s a bird. None of it is played as fantasy; it’s played as what a stressed brain does in a stressed place.

Yuts, the developer behind Geography of Robots, brought in a soundtrack from Gewgawly I, with Houston rapper Fat Tony featuring, and the music does something a lot of atmospheric indies fail at: it has a region. It sounds like it came from somewhere specific rather than from the drone-and-piano supply cupboard.

Where it fights itself

The minigames are the weak seam.NORCO periodically hands you a small mechanical diversion — a bit of combat-ish, puzzle-ish business — and these are fine, and they are also the least interesting five minutes on either side of them. They exist partly for rhythm and partly, I suspect, because an adventure game feels obliged to have verbs. The game is strongest when it trusts the cursor and the conversation. Compare whatHypnospace Outlaw does with the same problem: it never bolts on a verb, because it made its interface the verb, and it never has to change gear.NORCO changes gear, and you feel the clutch.

The bigger risk is legibility. This is a story that withholds, layers timelines, and expects you to assemble intent from fragments. Played across a few short sittings, whole threads can go slack. It rewards a couple of long evenings, and punishes the twenty-minutes-before-bed schedule that most of us actually have. That’s a real cost and worth knowing before you start, in the same wayPentiment is worth knowing about before you commit.

And the ending will annoy a certain kind of player. Fair warning. It’s an ending that resolves the emotional question and declines several of the plot ones.

The verdict

NORCO won the Tribeca Games Award in 2021, before release, off a slice — the first game to take that prize — and the festival juries were reacting to the right thing. It’s a genuinely literary game, in the narrow sense that its achievements are the achievements of prose: a sentence that lands, an image that won’t leave, an observation about people you recognise as true and hadn’t articulated.

What it does that a hundred other narrative indies don’t is refuse the easy consolation of resolution. Kay doesn’t solve Norco. Norco isn’t a mystery; it’s a condition. The game gives you a mind map for holding contradictions and then asks you to hold some.

It runs about five to six hours, which is the correct length, and it is short in the way a good novella is short. Play it on PC if you can — the cursor wants a mouse — though the 2023 console ports work fine and the Switch version is a decent bedside machine for it, schedule caveat notwithstanding.

Where next: if the mind map is what grabbed you,Disco Elysium is the fuller expression of the same idea. If it’s the place-as-character, takeCitizen Sleeper, which does economic precarity with dice instead of drainage canals and gets somewhere similar.

Spoilers below

The Superduck sequence is the moment the game shows its whole hand. Kay’s descent into a corporate-security theme-park apparatus, and the pivot from Southern Gothic to something closer to cyberpunk satire, is a swerve that shouldn’t hold and does — because the game has spent hours establishing that this town’s institutions are already absurd, so an absurd one arriving on schedule reads as continuity.

Catherine’s playable flashbacks are the structural masterstroke. Putting you inside the mother’s perspective, after hours of Kay assembling a picture of her from objects and other people’s accounts, means the game gets to do the thing prose does well and games usually can’t: show you that the picture was wrong in a way that isn’t a twist, just the ordinary gap between a parent and a child. You don’t learn Catherine had a secret. You learn she had a life.

And Blake. The search for Blake is the quest hook, and by the end the game has quietly demonstrated that finding him was never available, because what happened to Blake is what happened to the town — a slow dissolve into the machinery — and you cannot recover a person from an economy. That’s why the ending withholds. A resolution would have been a lie about how this works.

You wake on Erlin’s Eye, a half-derelict station built out of a scavenged shipyard, and you are legally a machine. Your mind is an emulation of a person who signed a contract; your body is a rented shell owned by a corporation called Essen-Arp, and it is decaying on schedule because the maintenance drug is a subscription you skipped by running away. That is the setup of Citizen Sleeper, made largely by Gareth Damian Martin as Jump Over the Age, published by Fellow Traveller, released 31 May 2022 on PC, Switch and Xbox, with a PlayStation port the following year.

The whole thing runs on five dice, and the dice are the best piece of design anyone has done with precarity.

What the pool actually models

Each cycle — the game’s word for a day — you roll a pool of dice. You spend them on actions: work a shift at the scrapyard, cook at the noodle stall, talk to someone, chase a lead. The number on the die sets the odds band for the action: a high die is likely to go well, a low die is likely to go badly, and a middling die sits in between. You allocate, you resolve, the cycle ends, you roll again.

The size of your pool is your Condition — the state of the body. Healthy, you get five dice. As the shell degrades you get four, then three, then two. Energy sits alongside it: eat and you function; go hungry and every die you roll is worse.

Read that back as a machine and see what it does.

It converts your health bar into yourscope of action. In almost every other game, low health means you might die. Here, low health means you can do fewer things per day. That is a different and considerably more accurate model of what being unwell actually costs a person: you do not lose a life bar, you lose Tuesday.

It makes randomness feel like circumstance rather than luck. A bad roll in a tactics game is a slap. A bad roll here is a morning where the work was there but your hands would not cooperate, and the game has already told you why — you are a Sleeper, the shell is failing, and this is what failing feels like from the inside. Same maths. Completely different meaning, because the fiction pre-explains the variance.

And it forces triage as the primary verb. Five dice, seven things worth doing. The station always has more work than you have hands, and every cycle you are deciding which relationship goes unattended so you can afford Stabiliser. Citizen Sleeper punishes you for having a finite number of hours, which is the only punishment the working poor ever actually receive.

Clocks, and why they hurt

The other half of the system is the clocks: circular progress trackers that fill as you commit dice, lifted openly from Blades in the Dark’s design vocabulary. Martin has never been coy about the tabletop debt.

The reason clocks work better here than a quest log would is that a clock isvisible partial progress on something you might abandon. A quest log says: this task exists. A clock says: this task is 60% done and it will still be 60% done in a month if you stop feeding it. Some clocks run backwards. Some are counting down towards you rather than towards a reward, and the sensation of watching a bounty-hunter clock fill while you cannot spare a die to deal with it is the single most stressful thing in the game.

That is what converts the dice from a puzzle into a life. A puzzle has a solution. A life has a set of clocks, all filling at different rates, and a fixed number of dice per day to distribute among them.

The nearest ancestor in games isPersona 5 Royal, which does the same thing with a calendar — a fixed number of afternoons and more people who deserve one than you have. But Persona’s scarcity is a scheduling optimisation you can solve with a spreadsheet, and the internet duly solved it. Citizen Sleeper’s scarcity has variance in it, which means it cannot be solved in advance and has to bemanaged in the moment. That is closer to how it feels.

The other ancestor is the board-game shelf. Dice allocation as a mechanism has been thoroughly worked out around kitchen tables — you roll, the number is your constraint, you place. Martin took a tabletop mechanism at the point where it was mature and asked what it would mean if the dice were a body. It is one of the few genuinely successful transplants of a board-game idea into a video game, and it works because the transplant carried thefeeling across rather than the rules.

The writing earns the frame

The prose is good enough to survive the amount of it there is, which is the minimum bar for a game where the dice mostly buy you paragraphs.

The station is a bureaucratic ruin: corporate remnants, a functioning market, an ecosystem of people who arrived for one reason and stayed for a worse one. The supporting cast are drawn with a restraint the genre rarely manages — a technician, a chef, a hauler, a kid, an AI — and none of them is a quest dispenser wearing a face. They have their own timetables. They leave.

Guillaume Singelin’s art gives the whole thing a warmth that argues productively with the material; the station is a nightmare of labour precarity rendered in soft, likeable lines, and that tension is deliberate. Amos Roddy’s score does the rest. The three free episodes — Flux, Purge and Refuge — that Jump Over the Age released across 2022 extend the story into the station’s refugee crisis and its politics, and they are the rare free DLC that changes what the base game was about.

Where it fights itself

The dice stop biting. By the late game, if you have played reasonably, you have solved your maintenance problem, your Condition is stable, and the pool is comfortably large. The precarity that made the first eight hours extraordinary becomes a formality, and the last stretch is a visual novel with a dice-rolling animation attached.

This is arguably thematic — the game is about escaping precarity, and escaping it should feel like something. I do not fully buy that defence. The best hours are the frightened ones, and a design that systematically dismantles its own central tension by hour twelve has traded its strongest hand for a narrative beat.

The second charge is smaller: the odds bands mean a low die is often simply wasted, and there are cycles where the correct play is to burn a die on nothing. A design that lets youdo something with your bad hours would have been truer.

The verdict

Citizen Sleeper does the thing I most want from a small game: it finds one mechanism, understands exactly what that mechanism means, and builds everything else in the frame to point at it. The dice are a body, a working week and a bank balance expressed as five cubes on a table, and for the first half of this game the sensation of placing them is as close as the medium has come to modelling what it is like to be short of everything at once.

It is on PC, Switch, Xbox and PlayStation, it takes ten to fifteen hours, and a sequel is on the way. Play the base game first and let the early cycles frighten you before you optimise them away.

For the other end of the same conversation — a game where the numbers on your character sheet are also the argument — readDisco Elysium, andNorco if what draws you here is a place that has already been ruined by capital and is still, somehow, inhabited.

Spoilers below

The endings are where the design’s honesty shows.

Citizen Sleeper offers several ways off the Eye and several ways to stay, and the game refuses to rank them. The route that gets you a body of your own, the route that puts you on a ship, the route that dissolves you into the station’s network, the route where you simply keep going — each is reached by having spent your dice on one set of clocks instead of another, over dozens of cycles, mostly without realising you were choosing.

That retroactive quality is the point. You discover, around cycle sixty, that the person you had been buying noodles for and the clock you had been quietly feeding had become the shape of your life, because those were the days you could afford. The endings are a summary of your scarcity.

The Essen-Arp material is handled with real discipline. The corporation never becomes a villain with a face. It is a legal position — you are property, you are in breach, and there is a hunter with a clock because that is what the contract provides for. When the game finally lets you address your status, the resolution is administrative, and it lands harder for it. A boss fight would have been a lie about how any of this works.

And the last thing, which I think is the game’s best line of thought: whatever ending you take, the station carries on. The market opens. The scrapyard needs hands. Someone else wakes up in a rented body with five dice and a subscription they cannot pay. Erlin’s Eye does not need you to have been there. It just needed a Sleeper.

Marissa Marcel made three films and none of them came out. Ambrosio in 1968, a Gothic thing about a monk, adapted from the Matthew Lewis novel that scandalised 1796. Minsky in 1970, a lurid detective picture. Two of Everything in 1999, a pop-star doppelgänger story after a thirty-year silence. She was the lead in all three. She vanished. The films were shelved.

None of this happened. Immortality, from Sam Barlow’s Half Mermaid, released on 30 August 2022 for PC and Xbox, hands you the surviving footage — clips, rehearsals, screen tests, behind-the-scenes offcuts, several hundred fragments in total — and gives you no index, no chapter list, no search box. It gives you one verb, and the verb is the whole game.

The match cut is a search query

Here is the mechanic. You are watching a clip. You click on something in the frame — a face, a lamp, a crucifix, a hand, a cigarette — and the game cuts you to a different clip, from a different film, a different decade, containing that thing. Then you do it again. That is the entire interface.

Consider what this actually is. It is a search engine whose query language isobjects in shot, and whose index you cannot see. You cannot ask for “1970, scene 14”. You can only ask for “somewhere else with a mirror in it”, and the game answers by throwing you thirty-one years across an archive that does not believe in chronology.

The system does three things at once, and this is why it is the best idea Barlow has had.

It makes browsing impossible, which forces attention. In search-box design — Barlow’s own Her Story, from 2015 — you are typing words you already suspect. Here you have tolook at the picture to find your next move, which means you are watching cinema the way a film editor watches cinema: scanning the frame for the object that will carry the cut. The game has trained a viewing habit into you within twenty minutes, and it did it by taking your index away.

It puts the connections in your head rather than the database. Two clips linked by a wine glass have no relationship the game has asserted. The relationship is one you built, because you clicked the glass. Every player’s Immortality is a different graph, and the game never has to author a single one of them.

And it makes the archive feelfound. An index implies a librarian. The absence of one implies the reels turned up in a lockup and nobody has catalogued them, which is exactly the fiction the game needs you to accept.

The films have to be good, and they are

This is the part that gets undersold. Immortality only works if three fake films, from three distinct decades, are individually convincing enough that you would watch them straight.

They are. Ambrosio is shot as a late-60s European art-horror piece, all shadow and religious hysteria, with the specific stiffness of a 1968 production that thinks it is being daring. Minsky has the greasy 1970 grain of a picture with a lower budget and a higher opinion of itself. Two of Everything is a 1999 slick thing, and the period detail extends to how the actors are beingdirected, which is a nuance almost nobody bothers with. The fictional directors — John Durick on the first and last, Arthur Fischer on Minsky — have distinguishable authorial tics, and you can tell whose set you are on before anybody speaks.

Manon Gage, as Marcel, is carrying a genuinely absurd load: she has to play a 21-year-old ingenue in 1968, the same woman hardening in 1970, and the same woman returning in 1999, across footage that you will encounter in random order and compare directly. She is superb. So is Charlotta Mohlin, whose work I will not describe above this line.

The production discipline behind this is what impresses me most as a piece of craft. Every clip has to be watchable cold, meaningful in context, and contain enough clickable objects to route you onward. That is three constraints on every frame of a feature-length shoot, times three films, and they were shot as real productions with period-appropriate technique.

The ancestor is Shattered Memories

Barlow’s obvious lineage is his own: Her Story (2015), Telling Lies (2019), a career built on giving players a pile of video and a way to interrogate it. Both of those games are search-box games, and Immortality is usually filed as the third one.

I think the real ancestor is Silent Hill: Shattered Memories, which Barlow wrote for Climax in 2009. That game’s actual idea was that it watched you back — the world reconfigured itself according to what you looked at and how you answered, and the horror was the implication of being profiled. Immortality is that idea with the profiling removed and the responsibility handed over. You are still being characterised by what you choose to look at. There is just nobody keeping the file.

The other ancestor is the CD-ROM crash of the mid-90s, and I say this as someone who watched it happen in real time. FMV died because the industry decided the videowas the game — press the right button, receive the next cutscene, and the interactivity was a toll booth on a film you were being shown. Barlow’s whole career is the correction: the video is thematerial, and the game is the apparatus you use on it. The scrub bar is the toy. Once you understand that, the entire genre reopens.

Where it fights itself

The randomness is a genuine cost. Clicking a recurring object gives you a clip from the pool of clips containing it, and the pool does not care about your progress. You will hit the same three fragments repeatedly while the one you need sits somewhere you have not thought to click. The game’s defenders call this serendipity. Some of it is; a fair chunk of it is churn, and the last stretch of a completionist run turns into pixel-hunting a frame for the object you missed.

The rewind mechanic — and I will keep this vague — is a second layer that a sizeable number of players never discovered unaided in 2022. The discovery rate on your central twist should probably not depend on whether the player idly held a button. It is a magnificent thing to find. It is also a design that has decided some of its audience will simply never see the game.

The verdict

Immortality is the most interesting thing anybody has done with video in a game, and it earns that by refusing the two easy versions: the interactive film where you press buttons, and the puzzle box where the video is a skin over a lock. The footage is the mechanism. You are the search algorithm. Every connection in your head was assembled by you out of raw material that was never sequenced.

It is on PC and Xbox, and it went to phones later, where it works better than you would expect because scrubbing is a touch verb. Give it a long evening with headphones and no walkthrough open. The moment when the archive starts answering back is worth protecting.

If the appeal is being handed a database and no instructions, readReturn of the Obra Dinn next, andLorelei and the Laser Eyes for the version where the archive is a building.

Spoilers below

The rewind is the game.

Holding the scrub backwards through certain frames peels the clip open and reveals another one underneath — footage that was never part of any production, in which two figures address the camera directly. The One and the Other One. An entity that has been inside Marissa Marcel, and the woman it displaced, both speaking from somewhere behind the film stock.

What makes this land is that the mechanic and the fiction are the same act. You have spent hours performing an intrusion — pulling apart other people’s work, watching rehearsals nobody meant you to see, freezing frames on faces between takes. The game’s answer is that something else has been doing exactly that, for much longer, and considerably better. The predatory viewer is the game’s actual subject, and it waited until you had become one before it told you.

The Ambrosio material is where the whole design justifies itself. A 1968 film about a monk destroyed by his own appetite, containing a hidden layer about a thing that consumes people to keep living, discovered by a player whose only verb islook closer. Three levels of the same idea stacked on one reel. The archive was built to carry that, and every other pleasure in the game is downstream of it.

The uncomfortable part, and the reason I keep going back to it, is Marcel herself. Every route through this game treats her as an object to be examined. That is what the interface permits. Barlow builds two hundred pieces of evidence that the film industry looked at this woman rather than at her work, hands you the tools to do the same thing, and then reveals that the looking was the horror. It is the neatest trap I have walked into in a decade.

The Tulpar is a freight ship crewed by five people, and its cargo is mouthwash. Pallets of it, being hauled across a lot of empty space by a haulage company called Pony Express, because a logistics decision somewhere required it. The ship crashes in the opening minutes. Nobody is coming.

Mouthwashing, from the small team at Wrong Organ and published by Critical Reflex, came out on PC on 26 September 2024 and takes about three hours. It is the most efficient horror game I have played in years, and the efficiency is the whole design rather than an accident of budget.

Three hours is a structural decision

Horror has a well-documented decay curve. The first hour of any horror game is the good one, because you do not yet know the rules; by hour six you have learned the monster’s aggro radius and you are playing a stealth game with a costume on. Alien: Isolation, which I admire enormously, spends its last third fighting this and losing. The dread converts into competence, and competence is the opposite of fear.

Wrong Organ’s answer is to finish before the conversion happens. Three hours is long enough to learn the Tulpar’s geography and short enough that you never master it. You are kept permanently in the first act’s emotional register — the one where you are still working out what kind of thing this is — and then the game ends while you are still in it.

That decision cascades through everything else. There is no combat, no inventory management worth the name, no crafting, none of the systems that games reach for when they need to fill hours they have already sold. Mouthwashing has almost no verbs. You walk, you look, you interact when the game lets you. The absence is load-bearing: a game with no way to fight back has no route to competence, and therefore no route out of dread.

The ship is the timeline

The Tulpar is small. You will walk its length dozens of times, and this is where the design does its cleverest work.

The game cuts between before the crash and after it, and it uses thesame corridors for both. The bridge you crossed as a functioning workplace is the bridge you cross as a tomb. Because you have physically walked it in both states, the comparison is stored in your legs rather than your head. Wrong Organ never has to tell you what has been lost, because you have the muscle memory of the version where it was fine.

This is spatial storytelling of a very old-fashioned kind, and it is enormously more effective than the audio-log approach that swallowed the genre after BioShock. A log is a thing a designer hands you. A corridor you have walked eighty times is a thing you own.

The crew are drawn with the same economy. Curly is the captain, wrapped in bandages after the crash and unable to speak. Jimmy is the co-pilot who is now in charge. Anya is the nurse. Swansea is the engineer, older than the rest, with the tired competence of a man who has watched several companies do this before. Daisuke is the intern, young and cheerful and the only person on board who seems to like his job. Five people, three hours, and every one of them lands. The writing gets there by giving each of them one thing they want and no way to ask for it.

The low-poly is an argument

Mouthwashing renders in a deliberately PS1-ish register: chunky polygons, texture warp, faces built from about nine triangles. The lazy version of this in 2024 is a nostalgia filter slapped over modern geometry, and there is a lot of it about.

Here it is doing something specific. Low-poly faces cannot emote precisely, which means the game must convey a state of mind through framing, sound and behaviour instead — and the player’s brain, given a face that will not resolve, fills in the worst available reading. Horror has known this since the Nostromo’s corridors were dark because the lighting rig was cheap. Ambiguity is free fuel.

It also solves the gore problem. A high-fidelity rendering of what has happened to Curly would be an endurance test and a certification headache. At this resolution it is a suggestion you cannot look away from, which is worse.Crow Country uses the same era for warmth and toyishness; Mouthwashing uses it to withhold.

What it is actually about

The horror here is administrative.

The mouthwash is the thesis. Five human lives are on that ship because a company decided the mouthwash needed to be somewhere else. When the crash happens, the cargo is revealed to be as worthless as it sounds, and the crew are left with a finite larder and a rescue that has no commercial reason to be dispatched. Swansea eventually drinks the stuff, because it contains alcohol, and that image — a man drinking the cargo that killed him — is the whole game in a frame.

What Wrong Organ understand, and what most workplace horror misses, is that the company never appears. There is no evil executive, no memo from the villain, no boss fight with a CEO. There is a haulage contract and five people inside it. The pressure comes from the shape of the situation, and the situation is entirely ordinary. This is where the real ancestor lives: the Nostromo of Alien (1979) was a working ship with a crew arguing about bonuses, and the reason that film has outlived a thousand monster pictures is that the monster was the second-worst thing on board. Mouthwashing skips the first monster entirely and keeps the bonuses.

The other ancestor is closer to home. This is a game about being trapped in a small space with people whose jobs are collapsing around them, and it belongs alongsideStill Wakes the Deep as evidence that the current wave of horror has worked out that labour is scarier than ghosts.

Where it fights itself

Two honest charges.

The non-linear structure is doing heavy lifting, and the cuts between timelines occasionally land as authorial rather than motivated — you can feel the writer choosing the moment to withhold. Most of the time this is fine, because the withholding is the point. Once or twice it reads as a magic trick rather than a choice.

And the interactivity is thin enough that a certain kind of player will reasonably ask what the game is doing that a short film would not. The answer is the corridors, and I think the answer holds — but it is a real question and it deserves a real answer rather than a shrug about “walking simulators”.

The verdict

Mouthwashing is three hours long, costs less than a takeaway, and will sit in your head for a fortnight. It works because every element of it is subordinated to one idea: keep the player in the state of not-yet-knowing, and then stop. There is no padding to defend, because there is nothing here that is not the point.

Play it on PC, in one sitting, at night, without reading anything else about it first — the discourse around this game gives away more than it realises. Then readSignalis for the other end of the same argument, where the low-poly frame is used for grief rather than dread.

Spoilers below

Jimmy is the reason to play this twice.

The game is narrated, structurally, by the crew member with the strongest possible motive to shape what you see. Jimmy caused the crash deliberately. Jimmy assaulted Anya. Curly, the captain, knew about the second thing and handled it by doing nothing at all — which is why the game’s most disturbing image is a man swaddled in bandages who cannot speak, being kept alive by the person he failed to stop. Curly is punished with a fate the game presents as unbearable, and it is unbearable precisely because you are asked to hold his complicity and his suffering at once.

The birthday sequence is what everyone remembers, and it is doing more than shocking you. Jimmy has spent the entire game constructing a version of events in which he is a man coping heroically with someone else’s disaster, and the cake is that fiction reaching its logical end: care and consumption performed as the same gesture. He is looking after Curly. He is also eating him.

On a second run, the earlier timeline stops reading as a workplace and starts reading as evidence. Every cheerful exchange has a second meaning. Daisuke’s enthusiasm becomes almost unwatchable, because you know what the ship is going to do with a person who is easy to like. And Anya’s scenes acquire a fury that the first playthrough has no way to register, because she is the only person on board who knows exactly what is happening and has already learned that saying so achieves nothing.

That is the trick, and it is a genuine one. The reveal does not rewrite the game. It rewrites you.

I loaded games off cassette on a C64 for most of my childhood, which means I spent a meaningful fraction of the 1980s listening to a tape deck screech at me for four minutes to find out whether the thing on the inlay card was any good. It usually was not. That is the part everybody forgets when they say the old games were better. The old games were a lottery, played at £1.99 a ticket, and the thrill was structurally inseparable from the odds.

UFO 50, released on PC on 18 September 2024 by Derek Yu’s Mossmouth, is a box of fifty tickets. It is presented as the complete catalogue of UFO Soft, a fictional developer who made games for a fictional 1980s machine called the LX System between 1982 and 1989. None of the games are real. All of the games are real. The distinction stops mattering about twenty minutes in, which is the trick.

The fiction is the design document

The team — Yu with Eirik Suhrke, Jon Perry, Paul Hubans, Ojiro Fumoto of Downwell fame, Tyriq Plummer and others — built the constraint before they built the games. One palette. One notional machine. One studio’s imagined career arc. Then they made fifty complete games inside it and dated them.

That constraint is doing three separate jobs, and only one of them is nostalgia.

The first job is coherence. Fifty unrelated minigames would be a Wario Ware. Fifty games from one imaginary studio have ahouse style — recurring mascots, sound palettes you recognise across a decade, ideas a designer clearly tried in 1983 and got right in 1987. Barbuta, dated first, is a deliberately obtuse thing that withholds almost everything and expects you to map it on paper. It is also the worst-reviewed game in the box by consensus, and it is placed first on purpose, because a studio’s first game is supposed to be the one where they had not worked it out yet.

The second job is permission. The fiction lets these designers ship a game that isrude. Real 1980s games did not explain themselves, did not respect your time, and frequently did not tell you the rules. Modern design has spent thirty years correcting that, mostly correctly. But a modern game that withholds is read as broken, whereas a 1984 game that withholds is read as a 1984 game. The frame buys the team the right to be genuinely unhelpful, and several of the best things in the box only work because of it.

The third job is the argument, which I will come to.

The good ones are properly good

The line that gets thrown around is that UFO 50 is “twelve great games and thirty-eight demos”. That is lazy. The distribution is real — there are entries here I bounced off inside five minutes and would not defend — but the hit rate is far better than any compilation cassette I ever owned, and the ceiling is higher than the pitch implies.

Mortol is the standout structural idea: a platformer where your stock of lives is the level’s building material, because each corpse becomes a step, a bridge, a switch held down. Dying is the verb. It is a genuinely publishable idea that would carry a full-price indie release on its own, and it is sitting in a box with forty-nine others.

Party House is a deckbuilder about hosting parties where the guests you want are the guests who might ruin it — a push-your-luck engine with a social skin that has no business being this tight. Grimstone is a full Western tactics RPG, hours long, with a party and a job system. Campanella is a physics flying game about momentum and patience. Velgress is a vertical climber built on the anxiety of rising death. Vainger is a Metroid-shaped thing with modular power slots. Golfaria is golf that grew a metroidvania. Night Manor is a point-and-click horror game with real dread in it.

Any one of those, polished up and released alone with a trailer, would have picked up coverage. That is the density we are talking about.

What holds them together is that the team understood which 1980s conventions wereload-bearing and which were merely damage. The obtuseness is kept, because obtuseness is what made a 1984 game a place you inhabited for a month rather than a thing you consumed in an evening. The genuinely broken parts are quietly fixed: the collision is honest, the inputs read on the frame you pressed them, the difficulty is hard in ways you can learn from. Anyone who has actually gone back to a beloved C64 title in the last decade knows how much of the misery was technical rather than intentional. UFO 50 keeps the intent and throws out the misery, and that editorial judgement — exercised fifty separate times — is the real labour in the box.

The cherry is the real design

The meta-layer is the part I keep thinking about. Finish a game and you get a Gift. Meet a harder, game-specific condition — a score, a challenge, a deeper completion — and you get a Cherry. The Gift says you saw it. The Cherry says you understood it.

This is a superb piece of engineering because it solves the compilation’s oldest problem. Every collection I owned as a kid had the same failure state: you play each thing for ninety seconds, decide, and never return. The Gift/Cherry split gives you two distinct reasons to stay, calibrated to two distinct kinds of player, and it puts the decisioninside each game rather than in a menu. You are never being asked to like all fifty. You are being asked to find out which three are yours.

The design ancestor is the high-score table on a machine in a chip shop, where the point was that someone else had already proved the number was reachable. UFO 50 rebuilds that pressure without a leaderboard, purely through the implied competence of a fictional studio.

The argument

Here is what the fifty fake games are actually arguing, and it is a better argument than the packaging suggests.

The claim is that the 1980s constraint producedvariety as a by-product of poverty, and that the variety was the good part. Nobody knew what a game was yet. A team of three had a machine with sixty-four kilobytes and no genre conventions to obey, so what came out was strange — golf with a map, a platformer made of corpses, a party sim, a mech game, a fishing thing — because nobody had yet worked out which of those were supposed to be commercially viable. The market answered that question in the 1990s and the answer narrowed everything.

UFO 50 stages a counterfactual: what if that decade had been run by people who already knew how to design? Same hardware ceiling, same palette, same absence of tutorials — with thirty years of accumulated design literacy behind the keyboard. The result is fifty games that feel period-accurate and are quietly smarter than anything the period actually produced, and the gap between those two facts is the thesis.

The honest ancestor of this whole object is the covertape and the budget label — Mastertronic, Codemasters, the magazine cassette blu-tacked to the front of Zzap! that I fed into a Datassette before I had read the review. Those tapes were the delivery mechanism for exactly this experience: a dozen unlabelled things, most of them bad, one of them yours forever. UFO 50 is that tape with the failure rate tuned down and the ambition tuned up.

Where it fights itself

The fifty-game shape has a fifty-game cost. There is no way to sample this efficiently, and the entries are not sorted by quality — deliberately, since the fiction requires a career arc rather than a greatest-hits. You will spend hours on things you do not like to reach things you love, and the game is perfectly comfortable with that. If your gaming time comes in ninety-minute slots after the kids are down, that friction is a real charge against it.

The other cost is that the deepest games here — Grimstone especially — are asking for the commitment of a standalone release while sitting behind a menu that implies a snack. Several of the best things in the box are structurally disadvantaged by their own container.

The verdict

UFO 50 is the most generous thing released in 2024 and one of the few games I would describe as an act of scholarship. It understands the 1980s as adesign condition rather than an aesthetic, which is why it earns the pixels in a way that a thousand pixel-art indies with a CRT filter never do. Buy it on PC. Understand that you are buying an argument with fifty pieces of evidence attached, that you will hate some of the evidence, and that at least two of the fifty will end up in your permanent rotation.

If the appeal here is the density of ideas per hour, the other 2024 indie worth your time isAnimal Well, andBalatro is the piece to read on what happens when one of these small, strange systems escapes the box and eats a year of everybody’s life.

Spoilers below

The catalogue has an internal continuity, and finding it is a genuine pleasure that the marketing sensibly left alone. Entries reference one another across the imagined decade. Mortol gets a Mortol II further down the timeline that takes the corpse-as-scaffolding idea and complicates it rather than merely enlarging it, which is exactly the move a real studio makes with a surprise hit. Mascots and sprites recur. Sound motifs carry between games years apart in the fiction.

The effect of that is stranger than a straightforward Easter-egg hunt. By the time you have played twenty of these, you have opinions about UFO Soft as acompany — which of their designers you rate, when they lost the plot, which 1986 experiment obviously came from the same person who made the 1983 oddity you disliked. You are doing criticism on a body of work that does not exist. That is an absurd thing for fifty games in a Steam release to achieve, and it is the clearest evidence that the fiction was the point rather than the wrapper.

Red Candle Games spent six years being known for two horror games and one international incident. Detention (2017) put White Terror-era Taiwan into a 2D side-scroller and got a Netflix series out of it. Devotion (2019) put a Taipei flat into first person, shipped with a piece of art mocking a head of state buried in a prop, and was pulled from Steam inside a week; the studio eventually re-released it through its own storefront in 2022. That is the reputation Nine Sols arrives against — a small Taipei team known for atmosphere, dread, and being difficult to buy.

Nine Sols, which came to PC on 29 May 2024 after a Kickstarter, is a 2D action game about deflecting. It is the least likely third act imaginable, and it is the best thing they have made.

The deflect is an investment, not attrition

Everyone will tell you Nine Sols is Sekiro in 2D, and everyone is right enough to be unhelpful. The comparison is worth making precisely, because the place where the two designs diverge is where Nine Sols becomes its own thing.

FromSoftware’s deflect inSekiro is attrition. Every parry you land pushes an enemy’s posture bar up and holds it there; the fight is a slow crowbar applied to a gauge, and the reward for perfect play is that the gauge stops draining. Deflecting is how yousurvive. Damage is the by-product of surviving well enough for long enough.

Nine Sols hands you a different contract. Your protagonist Yi carries the Foo Talisman: land a deflect, and you stick a charge to the enemy. The charge sits there. It does nothing on its own. You detonate it with a separate input, and that is where the damage lives. The parry is a deposit. The detonation is the withdrawal.

That single split changes the emotional texture of every encounter. In Sekiro you are pressing forward through defence. In Nine Sols you areaccruing — and the moment you notice you have three charges banked on a boss who is about to wind up something you cannot afford to interrupt, you have the specific, delicious anxiety of a man holding a full hand of chips at a table that might close. Greed becomes a mechanic. Do you cash out for a guaranteed chunk, or hold for one more deflect and risk eating the hit that wipes the ledger?

Then there is the Unbounded Counter, the charged answer to attacks marked in red that a normal deflect will not touch. It costs charge, it demands you hold the input through a window where you are committed, and it converts an unblockable into an opening. The red attacks are, in effect, the game asking whether you have been paying attention to the rhythm or merely surviving it.

Every one of those systems is a way of asking the same question: are you willing to standcloser than is comfortable? Nine Sols has no dodge worth the name in the FromSoftware sense; retreat is a losing strategy, and the game teaches this by making the rewards for proximity structural rather than cosmetic. It is generous with the lesson and merciless if you refuse it.

Why 2D is the right plane for this

There is a genuine engineering argument buried in Nine Sols, and it is the reason the Sekiro comparison flatters it.

Sekiro’s hardest problem is the camera. A deflect window measured in a handful of frames is a contract between the game and your eyes, and a 3D camera can break that contract without either party being at fault — a pillar intervenes, a boss steps behind you, the lock-on swings and you have lost the tell you needed. Every player who has bounced off a From game has a story that is really a camera story.

A 2D plane makes the contract enforceable. The tell is always legible. When Nine Sols kills you — and it will, repeatedly, and the second half is a step up that some players will find unreasonable — you know exactly which frame you got wrong. That legibility is worth more than any amount of tuning. It is the same reasonPrince of Persia: The Lost Crown felt so clean in the same year: constraining the axis is a feature.

The older ancestor here is not any Soulslike at all. It is the 8-bit fighting game. I spent a genuinely stupid portion of 1987 on International Karate + on a C64, and the thing IK+ understood — the thing Barbarian and its stablemates understood — is that a fight staged on a flat plane at a fixed distance is a conversation aboutspacing and timing, with no third dimension to hide the information in. Nine Sols is that conversation with thirty-seven years of animation budget attached. Yi’s sword has weight because you can always see the gap.

The frame, and the word “taopunk”

Red Candle coined “taopunk” for this, and the marketing instinct is a bit groan-worthy until you actually look at the place. New Kunlun is a Solarian colony rendered in hand-drawn art that puts Taoist cosmology on top of decayed industrial infrastructure, and thecombination is doing work rather than decorating. Cyberpunk’s usual grammar is Western corporate rot with a neon overlay. Nine Sols swaps the underlying philosophy out and the aesthetic reorganises itself around a different idea of what decline means.

Yi is one of the Nine Sols, awake after a long absence, hunting the other eight. The humans of New Kunlun are called Apemen and are treated roughly as you would expect a species to be treated when the people running the place regard them as raw material. The story is delivered in the Red Candle manner: patient, mostly environmental, unhurried about handing you the shape of it.

The build layer is jades — equippable modifiers you slot to shape Yi around your own bad habits. It is a light system by the standards of the genre, and I mean that as praise. The jades tune; they do not rescue. You cannot build your way out of failing to deflect, which is the correct decision for a game whose entire argument is that you should learn to deflect.

Where it fights itself

Two things.

The first is the difficulty step in the back half. Nine Sols is a game with a teaching curve of real elegance for its first stretch and a spike in its last that reads as the developers designing for the players who survived the first stretch. That is a defensible choice and a real cost, and anyone telling you the game is “fair throughout” is grading on the curve of having finished it.

The second is length. This is a fifteen-to-twenty-five-hour game with a platforming layer that is competent rather than inspired, and there are stretches of traversal between the combat set-pieces that exist because metroidvanias have traversal. The fights are where the design is thinking. The corridors between them are where it is filling.

Neither is fatal. Both are the kind of thing worth knowing before you commit twenty hours of your life, which is the only reason I raise them.

The verdict

Nine Sols is the rare homage that has an argument with its source. It took Sekiro’s central verb, worked out that the verb could be a currency rather than a gauge, and built a whole economy of greed on top of it — then staged that economy on a plane where you can actually see what you are doing. The result is a combat system that does something Sekiro does not: it makes you complicit in your own deaths. You did not fail to react. You held for one more charge.

That Red Candle got here from two horror games, via a delisting that would have ended a lesser studio, is the sort of career arc you do not get to see very often. Play it on PC. Give the first three hours the patience they ask for; the game is teaching you a verb, and it will not start speaking properly until you have it.

If you want the other end of the same year’s indie spectrum, the fifty-game argument of UFO 50 is worth your time next, andSignalis is the piece to read if what draws you here is Red Candle’s other register.

Spoilers below

Yi’s hunt has a shape you can see coming from a distance, and Nine Sols is comfortable with that. The revenge frame is a delivery mechanism for a question about what New Kunlun wasfor, and the answer — that the colony’s survival was engineered on top of the Apemen as a resource, with the Sols as the architects and Yi among them — recasts every fight you have had up to that point. You have been killing your colleagues over a decision you helped make.

The design consequence is the interesting part. The late bosses are the ones with the most personal claim on Yi, and the combat system’s greed loop lands hardest there, because the game has spent twenty hours training you to hold charges for one more deflect and the last fights are the ones where you most want it over quickly. The mechanic and the fiction end up asking the same thing: can you stand close to this a moment longer than is comfortable?

The Shuanshuan material — the small human boy Yi ends up responsible for — is the counterweight, and it is the reason the ending has any weight at all. Red Candle have always been better at the domestic scale than the cosmic one. Detention was a school. Devotion was a flat. Nine Sols is a colony, and the bit that works is still a kid asking for a story.