Split Fiction: Hazelight's Two-Player Grammar

Co-op games have a structural problem that most of them never solve: two players doing the same verb at the same time is only interesting for so long. Give both players a gun, or both players a jump button, and you’ve built a game that’s fun with a friend but not fun because of the friend — the second player is merely company. Hazelight Studios, Josef Fares’ team, has spent three games figuring out the actual answer, and Split Fiction, released in March 2025 through EA, is where the answer gets sharpest: give each player a different genre, running on the same screen at the same time, and make the seam between them the whole design.

Two writers, two worlds, one machine

The premise is a frame for the mechanic rather than a story worth dwelling on for its own sake: two novelists, Mio and Zoe, one writing science fiction and one writing fantasy, get trapped inside a machine designed to mine their imaginations, and find themselves thrown together into a shared simulation of both their manuscripts at once. Split-screen has been Hazelight’s format since A Way Out’s prison-break two-hander, but Split Fiction is the first time the studio has used the two halves of the screen to run two entirely different genres in parallel rather than two perspectives on the same event. It’s a rarer commitment to co-op-only design than most of the genre attempts — the kind of discipline Deep Rock Galactic’s co-op loop shows from a completely different angle, building every system around the assumption that a second player is always there rather than treating multiplayer as an optional layer over a single-player game. One half of a level might be a cover-shooter set on a cyberpunk skyway; the other, simultaneously, a dragon-riding fantasy chase — and the two halves are stitched together by a shared physical space both players are actually moving through together, a deeper join than a shared camera cut alone could manage.

The seam is the design

The genius of the format, and the reason it works where a lesser team’s version would just be two separate minigames glued at the middle, is that the two genres constantly have to solve the same spatial puzzle using completely different verbs. A chasm that player one crosses by grappling with a cybernetic arm has to be crossed by player two riding a conjured wind current, and the level geometry is built so both solutions have to land at the same physical point at the same moment for the game to let either of you through. That’s a much harder design problem than it sounds, because it means every set-piece is authored twice — once per genre — and then welded together so neither half ever waits idly for the other to catch up.

This is the throughline from Hazelight’s whole catalogue, refined rather than reinvented: It Takes Two’s central conceit was already “two different toolkits solving one shared puzzle,” expressed through the estranged-couple framing device and a new gimmick nearly every level. Split Fiction takes that same philosophy and removes the training wheels of a single shared genre, betting that two genuinely different worlds moving in lockstep is a more interesting collision than two variations on a domestic sitcom. The bet mostly pays off, because the genre-swap keeps a four-to-six-hour campaign from ever settling into a rhythm either player can get bored of.

Set pieces as the actual content

Hazelight’s games live or die on the density of new ideas per level, and Split Fiction’s hit rate is the highest the studio has managed. A single chapter might move from a side-scrolling beat-’em-up modelled loosely on old arcade brawlers, into a vertical platforming sequence built around gravity inversion, into a boss encounter that’s structured more like a rhythm game than a combat one — and each of those ideas gets exactly as long as it needs to land and not a beat longer. It’s the same design economy that made Ratchet & Clank: Rift Apart’s planet-hopping variety work: spectacle that refuses to overstay, delivered fast enough that the game never has to justify a mechanic’s existence beyond the ten minutes it’s actually fun.

The “side stories” — short optional detours themed around genuinely silly premises, a cow-riding rampage and a pig-racing gauntlet among them — exist purely as tonal relief valves between the main campaign’s more emotionally freighted chapters, and they’re honest about being throwaway in a way that keeps them from feeling like padding. A game this dense with new ideas needs somewhere to put the ones that don’t merit a full chapter, and the side stories are exactly that outlet.

The side stories also solve a pacing problem the main campaign can’t solve for itself: because the primary story is built from tightly authored, emotionally weighted set-pieces, it has nowhere to put a joke that’s purely a joke. Splitting that impulse off into consequence-free detours means the main campaign never has to awkwardly downshift tone mid-chapter, and the side stories never have to justify their own silliness against the emotional stakes of Mio and Zoe’s arc.

Camera as the third collaborator

Because two players are moving through visually distinct genres inside the same physical level, the camera has to do work no single-player game or symmetric co-op game ever asks of it: frame two simultaneous, tonally opposite scenes without either one reading as the background to the other. Hazelight’s solution is a dynamic split that isn’t a fixed vertical line down the middle of the screen but a boundary that shifts, widens and occasionally dissolves entirely into a single shared frame at moments the story wants read as unified rather than parallel. It’s an unglamorous piece of engineering that most players will never consciously notice, which is the correct outcome for camera work this load-bearing — the seam between two genres has to feel invisible even while the genres themselves stay completely distinct.

Built for two, and only two

The unfashionable thing about Split Fiction, and the thing worth stating plainly rather than treating as an asterisk, is that it has no meaningful single-player mode and no matchmaking for strangers in the way a live-service co-op shooter would offer. It requires a second person, physically or over the internet via the series’ now-standard friend-pass system that lets one purchaser’s copy grant a second, non-purchasing player free access. That’s a commercial risk in an industry that generally prices for solo play by default, and Hazelight has bet its entire output on it three games running. The bet keeps paying off because the format genuinely can’t be replicated solo — take the second player away and half of every puzzle’s geometry stops making sense, which is proof the co-op requirement is structural rather than a licensing gate bolted onto content that would work fine alone.

Where it strains

The dialogue between Mio and Zoe, meant to carry the emotional arc of two strangers becoming genuine collaborators, leans on well-worn beats about creative block and professional jealousy that the gameplay’s inventiveness has to work overtime to distract from. A handful of the fantasy-genre chapters lean on generic dragon-and-castle iconography that reads as noticeably less distinctive than the science-fiction half’s sharper visual ideas, suggesting the writing-and-art pipeline had more to say about Zoe’s imagined future than Mio’s imagined past. And because every level is authored as a two-genre pair, the game occasionally sacrifices depth in either genre for breadth across both — a cover-shooter section rarely gets to develop real tactical depth before the level moves on to the next idea.

The difficulty curve is also unusually flat by design — Hazelight has clearly optimised for two players of very different skill levels being able to finish the campaign together, which means neither the cover-shooter sections nor the platforming demands much precision from either player. That’s a deliberate accessibility choice rather than an oversight, and it’s the right one for a game whose entire audience is, by definition, playing with a second person whose skill level you can’t predict. It does mean a pair of experienced action-game players will find long stretches undemanding.

The real ancestor

The obvious lineage is Hazelight’s own back catalogue — Brothers: A Tale of Two Sons’ asymmetric twin-stick control scheme, A Way Out’s prison-break structure, It Takes Two’s toolkit-swap philosophy. The deeper ancestor, and the one that explains why the format lands as well as it does, is the arcade two-player brawler: Streets of Rage and Double Dragon’s insistence that two players occupying the same screen space, moving through the same level at the same pace, generates its own kind of drama regardless of what either player is actually doing mechanically. Split Fiction just replaces “both players punch the same enemies” with “both players solve the same geometry through different verbs,” which is the more sophisticated version of a design idea the arcade already knew was worth building an entire genre around.

The friend-pass distribution model is itself an inheritance worth naming: it echoes the arcade’s original economics, where a second player dropped a coin into the same cabinet rather than buying a separate machine, and Hazelight’s insistence on keeping that low barrier to entry across three consecutive releases is a rare case of a modern publisher preserving an arcade-era access pattern on purpose rather than by accident.

Spoilers below

The late-game reveal that the machine’s true purpose is to harvest and sell the writers’ unpublished manuscripts, rather than merely digitise them for the villainous publisher’s own catalogue, recontextualises every earlier chapter’s shift between genres as a literal theft in progress rather than a neutral simulation — the game has been showing you a robbery the whole time and only tells you so once escape is the only option left. The ending, in which Mio and Zoe co-author a final chapter blending both their genres rather than picking one, is the mechanical thesis stated as plainly as the story ever gets: the format was never about choosing a winner between science fiction and fantasy, only about proving the two could share a page.