The Bitmap Brothers Canon

The Bitmap Brothers were the first British developers to be photographed like a band. Mike Montgomery, Eric Matthews and Steve Kelly set up in 1987 in Essex, and within two years the style press was running features on them — leather jackets, sunglasses, the lot. The industry found this hilarious and slightly embarrassing, and a lot of people concluded that a studio marketing itself that hard must be hollow underneath.

They were wrong, and the reason is worth pulling apart. The chrome logos, the Mark Coleman artwork, the licensed dance records on the title screens: all of that was doing legible design work. A Bitmap game announces its rules through its surface. The metal tells you the game will be heavy and unfair. That’s a contract, and they kept it more often than most studios keep theirs.

Ten games, ranked by nothing, sorted by what each one was actually for.

The Melbourne House years (1988)

Xenon (1988). A vertical shooter whose one idea is the transform: your ship switches between a fighter and a tank, and the tank can take ground routes the fighter can’t. The switch costs you a beat of vulnerability, so the mechanic is a gamble rather than a loadout. Rough around the edges and clearly a first game, and the transform is a better verb than nine-tenths of the shooters that followed it. The thing to notice is that the two forms have different collision profiles against the same level, so the game is quietly asking you to read each corridor twice — once as a flier, once as something with tracks. That’s a lot of thought for a 1988 budget-adjacent shooter.

Speedball (1988). The original is a harder, colder thing than its sequel — a two-team future sport with a ball, a pit and no referee. The design question it asks is what happens to a sport when you delete the rulebook and keep the scoreboard. The answer turned out to need one more iteration.

The peak (1989–91)

Xenon 2: Megablast (1989). Developed by The Assembly Line to a Bitmap design and published through Image Works, and the game everybody remembers for the soundtrack — Bomb the Bass’s “Megablast” as the title music, which in 1989 was genuinely startling. The mechanical hook is the shop: you earn currency, and between levels you buy weapons from a merchant. Attaching an economy to a shoot-’em-up meant the difficulty curve became partly your responsibility, which is the ancestor of every roguelite shop screen currently in the charts. The scrolling backwards is the divisive part; I’ve come round to it, because being able to retreat over ground you’ve cleared makes the level a space rather than a corridor.

Cadaver (1990). An isometric dungeon game with a physics-ish puzzle logic — objects have weight and placement, and most problems are solved by moving something heavy onto something else. It’s the least fashionable Bitmap game and the most patient, and it sits squarely inside the isometric canon without ever getting mentioned in it.

Speedball 2: Brutal Deluxe (1990). The one. A nine-a-side future sport with a league, a transfer market, an armour economy, and a scoring system that awards points for injuring opponents as well as for goals. That last rule is what makes it a great design: violence has a number attached, so brutality becomes a legible strategy you can plan rather than a texture. Add the star-domes, the multipliers and the bounce-walls, and you have a sport where a losing team can build a comeback out of pure malice. Still the best 16-bit sports game there is. The full argument, and why 16-bit sports games were better systems generally.

Gods (1991). Richard Joseph’s audio, a Nation 12 theme in “Into the Wonderful”, and a platform-puzzler with something almost nobody noticed at the time: an adaptive difficulty system. The game watches how you play and adjusts enemy behaviour and item drops accordingly — die repeatedly on a room and it eases; carve through it and the game starts sending harder patrols and better loot. In 1991, when the industry’s entire vocabulary for difficulty was a menu with three words on it, that is a startling piece of engineering, and it lands years before the adaptive-encounter systems people credit to much later games. The stated design goal — “to entertain” — was printed on the box, and the game does the work quietly underneath. It’s also brutally slow if you fight everything, because Gods wants you to think about the room before you enter it.

Magic Pockets (1991). The odd one out: bright, cartoonish, a kid with pockets full of weapons, and a Betty Boo theme tune. The Bitmaps doing whimsy is a strange sight, and the game underneath is a competent collect-’em-up rather than a great one. It belongs in the canon as evidence that the house style was a choice, since they could clearly switch it off.

The Renegade years (1993–96)

The Chaos Engine (1993). Six mercenaries, a Victorian England ruined by a machine, and a top-down run-and-gun built for two players sharing one purse. The shared-money design is the argument: you and your partner upgrade from the same pot, so every purchase is a small negotiation, and a greedy co-op partner is a mechanical problem rather than a social one. The level design is a maze of gates and switches that rewards splitting up, which pulls against the shared economy in a way that keeps the tension live for the whole game. The steampunk run-and-gun in full.

The Chaos Engine 2 (1996). Arrived years late into a market that had moved to 3D, and turned the co-op into competitive play. The idea was reasonable and the timing was fatal. Worth knowing about; hard to recommend as an evening.

Z (1996). A real-time strategy game with no base-building economy — you take territory by capturing flagged zones, and the zones produce units automatically. Deleting the resource-gathering layer that Dune II had installed as compulsory was a real design argument, and it made Z a game about map pressure and timing rather than about mining. The FMV cutscenes with the swearing robots got the attention. The territory model is the part worth stealing, and you can see its descendants in every point-capture RTS since. Dune II wrote the template Z was arguing with.

The bit nobody talks about: they bought their own printing press

In 1991 the Bitmap Brothers co-founded Renegade Software with Rhythm King, the dance label — and that is the most consequential thing the studio ever did. Developers of the period signed to publishers who owned the masters, set the schedule and took the margin. Renegade meant the Bitmaps published themselves, and then published other people: Gods, Magic Pockets and The Chaos Engine came out through it, and so did Sensible Software’s Cannon Fodder and Sensible World of Soccer.

Two of the best British studios of the era were releasing through a label owned by one of them. That’s the structural reason the Bitmaps could spend two years on a game while everyone else shipped annually, and it’s the reason their catalogue is small and their hit rate is high. The leather jackets got the column inches. Owning the release schedule is what actually bought the quality.

What the style was for

The obvious read is that the Bitmaps were an art department with a programming team attached. The better read is that they were unusually good at communicating a system before you played it. Look at a Speedball 2 sprite: heavy, armoured, low centre of gravity. It moves the way it looks. Look at the Chaos Engine’s mercenaries, each with a silhouette that tells you their speed. That’s the same trick Sensible Software pulled from the opposite direction — tiny sprites, instantly readable — and both studios were solving the same problem of getting rules into a player’s head without a tutorial. Style as substance, taken apart properly.

There’s a limit to the defence. Xenon is stiff. Magic Pockets is slight. Chaos Engine 2 and Z arrived after the ground had shifted, and the studio’s habit of announcing a game and then perfecting it for two years stopped being charming once the market started turning over annually. The Bitmaps were built for a world where a small team could hold a technical edge for a couple of years, and that world ended around 1994.

What survives is the discipline. Four of these — Xenon 2, Speedball 2, Gods, The Chaos Engine — are games where one clean mechanical idea carries the whole thing, and the styling exists to explain the idea. Set that against Psygnosis, who were selling the same era on presentation and often had nothing underneath, and the distinction gets sharp. Both studios made your Amiga look expensive. One of them made the machine worth owning.

Where to play them: most of the catalogue is available through the Amiga re-release collections and mini-console packages, and Speedball 2 in particular has been ported to roughly everything with a screen. Take the original Amiga versions where you can get them — the conversions varied wildly, and the console ports mostly lost the weight.