Hunter: The Open World in 1991

In 1991, on an Amiga, you could stand on a beach, look at a jeep, get in it, drive it inland until you found a helicopter, get in that instead, fly to another island, land, steal a bicycle, and ride the bicycle to a building you had been asked to blow up — and the game had no opinion about any of it. Hunter did that. Paul Holmes wrote it, Activision published it, and the industry proceeded to spend the following decade rediscovering the idea one component at a time.

I want to make the case for Hunter carefully, because the temptation with a game like this is to overclaim. Here is the modest version, which is remarkable enough on its own: in 1991, on hardware with no floating-point unit and no texture mapping, Hunter shipped a working implementation of a design pattern that would define the following twenty years. It is worth looking at what it got right on the first attempt, and equally worth being honest about what it never managed at all.

What is actually there

The world is filled-polygon 3D — flat-shaded, untextured, running at a frame rate that is best described as sincere — and it is made of islands you can move between. There is a third-person camera behind your soldier. There are buildings, trees, walls, hills, water. There are vehicles standing around in the world as objects, and the interaction with all of them is the same: get close, get in, drive it. Jeeps, tanks, boats, helicopters, hang-gliders, bicycles, motorbikes. Some of them handle appallingly. All of them work.

There is a day/night cycle, which does exactly what a day/night cycle should do and almost never does: it changes how you play. Night reduces your visibility and the enemy’s, which turns darkness into a tactical resource rather than a lighting filter. You can wait for it. A game in 1991 that rewards patience with a clock is doing something that a great many contemporary open worlds, with their vastly more expensive skies, still fail to do.

You get missions as a list, and you do them in whatever order you like, from wherever you happen to be, using whatever you can find. There is no gating. The game’s structure is a set of objectives and a landmass, and the join between them is your problem. That is the whole design, and in 1991 it was close to unprecedented as a commercial product.

The verb that matters

Every open world lives or dies on one question: what does the player do with the space between the objectives? Hunter’s answer is traversal, and it commits to it harder than most modern games dare.

The commandeer mechanic is the key. Because every vehicle in the world is enterable and none of them are yours, the world stops being scenery and becomes inventory. You look at a hillside and see a route, a boat, a place to land. The mental model the game builds is the same one Grand Theft Auto III would build a decade later, and it is built by exactly the same mechanic: universal vehicle access with no ownership. That single rule converts a landscape into a system, because it means everything in view is potentially a tool.

The physics are the other half. Boats behave like boats, badly; the hang-glider requires thermals and patience; the helicopter is a genuine skill check that will kill you repeatedly before it becomes the best thing in the game. Each vehicle is a different verb with a different competence curve, and the world is large enough that choosing wrong costs you real time. That is a proper systems economy, and it emerges from a fleet of objects rather than from a designed progression.

Hunter also lets you be on foot in the same space, which sounds trivial and is the hardest part. Most vehicle games of the era put you in a cockpit and left you there. Keeping the soldier as the persistent unit — the thing that gets in and out, that swims, that has to eat — is what makes the world continuous. There is no transition. You are always the same object in the same space.

The argument about GTA

So: is Hunter an ancestor of Grand Theft Auto III? I want to argue this from the record rather than assert it, because the internet’s version of this claim is usually a lot louder than the evidence supports.

What can be said with confidence is that Hunter demonstrates, in 1991, a working combination of features that would later be treated as GTA III’s signature: freeform third-person movement through a contiguous 3D world, universal vehicle commandeering, non-linear mission order from a hub, and a simulated environment that runs whether you are looking at it. That combination existing that early is a matter of public record, and it is genuinely remarkable.

What cannot be said is that DMA Design took it from Hunter. There is no documented line of influence I am aware of, and the fact that both games are British is a coincidence with a large denominator. Ideas of this kind arrive when the hardware makes them cheap, and by 2001 the hardware had made them cheap for everybody. The honest claim is convergence: Hunter shows that the design was thinkable, and buildable, a full decade before the market was ready to reward it.

That is a more interesting claim than influence, actually. It means the open world was never waiting on an idea. It was waiting on the polygon budget that would let the idea look like something people wanted to buy. Hunter is the proof, because it has the idea and manifestly lacks the budget.

Why nobody cared

Hunter’s problem is legibility. In 1991 a screenshot of it looks like nothing — flat grey shapes on a flat blue background — and the Amiga market ran on screenshots in magazines. The pleasures it offers are structural, they take twenty minutes to arrive, and none of them photograph.

Meanwhile the games selling next to it were selling on exactly what Hunter cannot do. Shadow of the Beast moved units on the strength of a parallax scroll you could see from across a shop, and that was the market’s actual currency. A game whose selling point is “the missions can be attempted in any order” had nothing to put in the box art.

The other problem is that Hunter is genuinely awkward. The controls are of their era and their era was not kind. The frame rate makes precision difficult. The combat is thin — shooting in Hunter is a formality, and the game knows it, which is why the interesting content is all about getting places. A design this far ahead of its hardware pays for it in feel, and feel is what most players judge in the first five minutes.

There is a third reason, and it is the one that generalises. Hunter’s core pleasure is a negative — the absence of a rail. It gives you a world and declines to tell you what to do with it, and in 1991 the audience had no vocabulary for that absence and largely read it as the game failing to have a structure. The freedom that would later be the entire selling point of a genre looked, at the time, like a missing feature. Every review of Hunter had to invent the concept before it could praise it, and reviews rarely have room to invent concepts.

The British 3D tradition it belongs to

Hunter did not appear from nowhere. It sits in a specific and rather glorious British lineage of programmers doing impossible geometry on home computers: Elite with its wireframe galaxy in 32K, The Sentinel with its solid landscape and its single terrible idea, and Midwinter, which a year before Hunter had already put a huge traversable 3D world and a survival economy on the Amiga. Midwinter is the closest relative and the fairer comparison — Holmes was working in a tradition that Mike Singleton had already opened.

What that lineage shares is a willingness to spend the entire machine on space and let everything else be crude. It is the opposite instinct to the one that produced the Amiga’s famous games, and it is the one that turned out to be right about where the medium was going.

What it can teach a 2026 open world

The uncomfortable thing about going back to Hunter is how much of the modern genre it makes look like decoration. A contemporary open world spends its budget on density: things to see, markers to clear, a landscape upholstered until no minute is unaccounted for. Hunter has an island, some vehicles and a list. The space between objectives is empty, and the emptiness is what makes the traversal mean anything, because a journey through a landscape that has been pre-loaded with entertainment is a corridor with better wallpaper.

Hunter also gets the vehicle question right in a way that money has since made harder. Because the fleet was crude, every vehicle in it had to justify itself mechanically, and each one is therefore a genuinely different proposition. When the polygon budget arrived, the industry mostly spent it on making vehicles look correct and handle pleasantly, which is a lower ceiling than making them feel distinct. The helicopter in Hunter is difficult. Difficulty is why it is the best thing in the game.

Where to play it

Amiga, under emulation, with a manual open — Hunter is from the era when the manual was part of the software, and it will not explain itself. Give it an hour, ignore the shooting, and go and find the hang-glider. The moment the game lands is the moment you realise you have crossed to an island you were not sent to, in a vehicle nobody gave you, for no reason. That feeling was already a decade old when the games built on it arrived, and it is thirty-five years old now.