Why Bad Graphics Work in Some Games

Why bad graphics work in some games is a question that misframes the design choice it’s asking about. “Bad” suggests failure. What’s actually happening in A Dark Room, Universal Paperclips, ASCII roguelikes, and the Chrome Dino game is deliberate abstraction — a refusal to render details that the player’s imagination renders better. The format isn’t a budget compromise; it’s a design strategy with its own grammar.

What “bad graphics” actually means

“Bad graphics” in common usage describes any game that doesn’t aim for photorealistic or high-fidelity art. The category includes deliberate minimalism (Universal Paperclips), retro pixel art (Chrome Dino, Celeste), ASCII rendering (NetHack, Dwarf Fortress), text-only presentation (A Dark Room, Zork), and constrained palettes (Downwell, Return of the Obra Dinn). Most of these are not bad graphics; they’re chosen art directions that look “bad” only by the photorealism benchmark.

The real distinction worth drawing is between underdesigned graphics (where the budget ran out) and underspecified graphics (where the designer chose to leave details to the imagination). Underdesigned art is bad. Underspecified art is often the best design choice the game could make.

The imagination gap

The central mechanism of underspecified graphics is what cognitive scientists sometimes call the imagination gap. When the game refuses to render a detail, the player’s mind fills it in. The filled-in version is often more compelling than any rendered version would be, because the player’s imagination is tuned to their own preferences, fears, and aesthetic sense.

This is why text adventures and ASCII roguelikes can produce horror, awe, and tragedy at a scale that high-fidelity 3D games struggle to match. A Dark Room’s late-game narrative reveals are devastating partly because the player has been imagining the world for hours and has invested in that mental image. Replacing the text with rendered cutscenes would diminish the effect, not enhance it.

The same principle operates in other media. Radio drama produces vivid imagery without any visual at all. Prose fiction renders worlds the reader fills in. Game design that uses graphics sparingly is borrowing from those older formats.

A Dark Room: text and counters

A Dark Room (2013, Doublespeak Games) is the cleanest demonstration of the principle in browser games. The game has no graphics — just text descriptions and numeric counters. The room is described, not drawn. The stranger who arrives at the door is described, not drawn. The village that builds itself around the room is named, not visualized. The player imagines all of it.

The game’s late-stage narrative reveals work because of the abstraction. When the player learns what their actions have actually meant, the recontextualization lands harder because they had been constructing the world themselves. A graphical version would have constrained their mental model and softened the impact.

Universal Paperclips: numbers as the world

Universal Paperclips (2017, Frank Lantz) takes the abstraction further. The game’s world is entirely numeric — counters for paperclips, wire, money, trust, probes, available matter. There are no characters. There are no environments. There are no animations beyond updating numbers. And yet the game produces a real sense of escalation as the player progresses from making paperclips one at a time to converting solar systems.

The numbers carry the experience because the player projects scale onto them. A counter that reads “10,000,000,000” feels different from one that reads “100” not because the game has rendered the difference but because the player has internalized what the difference means. The game uses the player’s existing intuitions about scale rather than trying to render scale visually.

ASCII roguelikes: depth without art

NetHack (1987), Dwarf Fortress (2006), and Caves of Qud (2015) are deep simulation games rendered in ASCII characters. A ‘@’ is the player. A ‘D’ is a dragon. A ‘#’ is a corridor. The art surface is, by photorealism standards, nonexistent. The systems underneath the ASCII are some of the deepest in any game.

The ASCII format does two things at once. It reduces the development cost of rendering to near zero, which lets the design budget go to systems. And it forces the player’s imagination to construct the world, which makes the player’s investment in any given playthrough feel personal. Dwarf Fortress players famously construct elaborate mental narratives around their dwarves, which would be harder to do if the game showed them a specific rendered dwarf face.

The trade-off is accessibility. ASCII roguelikes have a steep learning curve precisely because the player has to learn the visual vocabulary before play can begin. Most modern descendants (Caves of Qud, Cogmind, Brogue) add optional tile sets that retain the ASCII grid structure with rendered icons, which lowers the entry cost without sacrificing the format’s depth.

Pixel art as the durable form

Pixel art is the most commercially successful form of “bad graphics” in modern games. Celeste, Stardew Valley, Hyper Light Drifter, Hollow Knight (sprite-based with some pixel work), Undertale, and the Chrome Dino runner all use deliberate pixel art for aesthetic reasons. The form looks dated by photorealism standards but doesn’t age, because it was always pixel art on purpose.

The aesthetic logic of pixel art is the same as the imagination-gap logic: the game shows a stylized representation, the player’s brain fills in the implied detail. A pixel-art tree is recognizable as a tree without rendering any individual leaf. A pixel-art character has personality in the silhouette without needing facial animation. The economy of the form is part of its expressive power.

Chrome Dino specifically uses pixel art that mimics early-1980s arcade sprites. The dinosaur, the cactus, the pterodactyl, and the moon are all rendered in small palettes with no anti-aliasing. The style fits the Easter-egg context (a small playful thing on an error page) and reads as deliberate art direction rather than budget compromise.

The budget reallocation

One under-discussed effect of underspecified graphics is that it frees development budget for everything else. A studio that doesn’t have to render a high-fidelity world can spend that budget on systems, writing, music, and design iteration. A Dark Room’s narrative depth is partly funded by the absence of art assets. Universal Paperclips’ system depth is partly funded by the same.

This is the practical argument for the format. Most small studios don’t have the budget for high-fidelity art. The choice isn’t between “good graphics” and “bad graphics” — it’s between rendered art that competes badly with bigger budgets and stylized art that competes on its own terms. The smart move is usually the stylized one.

Where the format fails

Underspecified graphics fail when the game’s experience depends on the missing details. A combat-heavy game needs animation to convey kinaesthetic satisfaction; pure text won’t carry it. A platformer needs visual feedback on collision and timing; ASCII would frustrate. The format works for genres where the experience lives in systems, writing, and player imagination — narrative incrementals, roguelikes, simulation games — and fails for genres where the experience lives in the rendered world.

The format also fails when designers reach for it without commitment. A text adventure with a few half-hearted screenshots feels neither like a text adventure nor like a graphical game. The best examples (A Dark Room, NetHack, Universal Paperclips) commit to the constraint completely. For more on the textual tradition specifically, see our roundup of best browser text adventure games.

Frequently asked questions

Are text-only games still relevant in 2026?

Yes. A Dark Room, Universal Paperclips, and modern interactive fiction (the Choice of Games line, Choice of Robots, Twine games on itch.io) all sustain active audiences. The format has a smaller mainstream footprint than 3D games but a durable one.

Why do ASCII roguelikes still attract players?

The depth of systems compensates for the visual abstraction. Dwarf Fortress simulates hundreds of dwarves with individual personalities and histories; NetHack has interaction rules between every item and creature in the game. The ASCII surface is the cost of admission to systems that high-fidelity games can’t afford to build. The Roguelike Wikipedia entry covers the lineage.

Is pixel art “bad graphics”?

No, it’s a deliberate style with its own aesthetic logic. Pixel art borrows the constrained palettes and sprite-based rendering of 1980s arcade games and uses them for expressive effect. Celeste and Hollow Knight are the modern reference points.

Why did A Dark Room have no images?

Michael Townsend, the developer, chose the text-only presentation as a design strategy rather than a budget constraint. The format lets the player’s imagination construct the world, which makes the late-game narrative reveals more impactful.

Can a game with bad graphics be a hit?

Yes. A Dark Room topped App Store charts in eighty-plus countries. Universal Paperclips was covered by The Atlantic and Wired. Dwarf Fortress has a paying Steam audience after decades as a free ASCII game. The format is commercially viable when the underlying design is strong.

The takeaway

Why bad graphics work in some games comes down to abstraction as design strategy. Underspecified art lets the player’s imagination fill the gaps, frees the development budget for systems and writing, and produces a different kind of immersion than high-fidelity rendering can. The format isn’t a compromise; it’s a tool. For a small browser game whose deliberate pixel art belongs to the same family — minimal palette, no animation beyond the necessary, sprite work on purpose — the Chrome Dino game at the top of this site is the everyday example. It looks “bad” in the same way A Dark Room looks “bad,” which is to say it doesn’t look bad at all.