ADOM (Ancient Domains of Mystery)

Start with the date, because the date is the whole story: 1994. ADOM has been in active development since Bill Clinton’s first term. Thomas Biskup — a German computer scientist who started the project as a student — has been iterating on the same game for longer than most studios have existed. The Steam release on November 16, 2015 wasn’t a launch. It was a formal introduction of something that had already spent two decades metastasizing in the freeware roguelike scene, where it sat alongside NetHack and Angband as one of the three pillars of the genre.

That lineage matters, because it explains the shape of the thing. ADOM is not a modern roguelite with a meta-progression tree and a 20-minute run length. It’s a full turn-based RPG in the classical sense: you build a character from a deep pool of races and classes via point-based generation, you walk out into the world of Ancardia, and you try to stop the forces of Chaos from unmaking reality. A single serious attempt can run dozens of hours. Hundreds of locations. Over 400 monster types. Tens of thousands of items. An actual plot with multiple distinct endings that materially change the fate of the world — a genuine rarity in a genre that usually treats narrative as flavor text bolted to a random number generator.

Biskup didn’t do the Steam version alone. Jochen Terstiege, Zeno Rogue, Krzysztof Dycha, and Lucas Dieguez round out the credited team, and the collaboration shows in what got added: a Story mode with save/restore for people who won’t accept permadeath, a Challenge mode with weekly seeded competitions for people who demand it, an Exploration mode with wishes for people who just want to see the museum, selectable difficulty, ghosts left behind by other players’ deaths, 70+ achievements, trading cards, and a full graphical tileset you can toggle against the original ASCII at will.

The market positioning is unusual and, frankly, honest. This is a $14.99 premium artifact competing against free-to-play roguelites with slot-machine reward loops and Hades-grade production values. It wins on exactly one axis — depth — and it wins on that axis by a wide margin. That is a cult-consensus proposition, not a mass-market one, which is precisely what it should be.

• Corruption as a Risk/Reward Engine: Power drawn from Chaos slowly mutates and deforms your character. This isn’t a debuff meter. It’s a design thesis — every strong option carries a permanent physical cost, and the game’s most interesting decisions come from deciding how much of yourself you’re willing to lose.
• Three Modes, Three Audiences: Story (save/restore, forgiving), Challenge (permadeath, weekly seeded competition against the world), Exploration (wishes, tourism). One purchase, three genuinely different games. This is the smartest thing the Steam release did.
• Systemic Density: Melee, ranged, and spell combat, plus skills, crafting, alchemy, religion and sacrifice, and a corruption system that touches all of it. The systems interlock rather than sit in parallel silos — sacrificing at an altar affects your god, which affects what you can survive, which affects how much corruption you can afford.
• Dual Rendering: Full graphical tileset or traditional ASCII, toggled freely. Not a compromise between old and new. Both, on demand.
• Player Ghosts & Weekly Seeds: Asynchronous multiplayer done with restraint — you find the remains of other players’ failures, and everyone races the same seed weekly. Community without lobbies.

The Gameplay Loop

There isn’t one. Not in the modern sense, anyway, and that’s the point worth understanding before spending money.

Contemporary roguelikes are built around a tight, legible loop: enter run, acquire build, die, unlock something, repeat. The loop is the product. ADOM’s loop is closer to a campaign — you’re managing a single character across an entire world map, a dozen quest lines, towns you return to, dungeons that persist, and a corruption clock ticking in the background that gives the whole thing a slow-burn dread most games in the genre never attempt.

That corruption clock is the game’s standout mechanic and deserves a hard look. Chaos grants power. Chaos also mutates you — permanently, visibly, cumulatively. You grow extra limbs, lose senses, become something that can’t wear its own armor anymore. The mechanical consequence is that ADOM’s difficulty curve runs in two directions at once: you get stronger, and you get worse, and the skill of the game is arbitraging that gap. Most RPGs give you a power fantasy with a resource cost. ADOM gives you a power fantasy with an identity cost. Thirty years on, few games have matched it.

The systems density is real and not padding. Alchemy, crafting, religion, sacrifice, skill training — these connect. Your god’s disposition is a survivability resource. Your survivability determines your corruption tolerance. Your corruption tolerance determines which power spikes you can accept. Pull one thread and three others move. That’s the mark of a designer who’s had thirty years to sand down the seams between subsystems, and it’s why the game still holds up against titles with a hundred times its budget.

Onboarding Friction Is the Central Problem

Now the bad news, and it’s substantial.

ADOM is opaque by design and hostile by accident. The learning curve is brutal, the game leans on unmarked traps and unsignposted ambushes to kill characters players have invested ten hours in, and the practical solution the community has arrived at is read the wiki. Not “read the wiki for optimization.” Read the wiki to know that a specific room type will end you.

There’s a defensible version of this argument — discovery through death is a genre tradition going back to Rogue itself, and permadeath means nothing if the world is fully legible. But there’s a difference between a game that withholds information and a game that punishes you for not having consulted an external document. ADOM lands on the wrong side of that line more often than it should. When the optimal play pattern involves a second monitor, the design has outsourced part of itself.

The interface compounds it. Dozens of keyboard shortcuts, layered menus, and an information architecture that assumes you already know where things live. The Steam release modernized the UI meaningfully, but modernizing a 1994 interface gets you a well-organized 1994 interface. The command vocabulary is still enormous. The first hours are less “learning a game” than “learning a text editor.”

Who This Actually Rewards

Push past that wall and the payoff is rare. The reason this game sustains a devoted following despite airing every one of these complaints is that little else offers this much game. The variety of races, classes, skills, items, and quests isn’t a bullet point — it’s the reason a character build in hour forty still feels like it’s revealing new options. The Chaos plot gives runs a destination rather than a score. Multiple endings mean the finish line isn’t binary.

And Biskup’s update policy — buy once, free updates, forever — is the kind of thing that’s become quietly radical. There’s no season pass. No cosmetic store. The game you buy is the whole game, and it’s been getting bigger for three decades. The tradeoff is that updates arrive on geological timescales. Long silences between patches are the cost of a small-team passion project, and occasional bugs persist longer than they would in a studio product.

The aesthetic direction is best described as functional. The tileset works — it’s readable, it does the job of translating 400+ monster types into distinguishable sprites, and it lowers the entry barrier for anyone who can’t parse a screen of letters as a dungeon. It is not attractive. The graphics are dated even with tiles on. There’s no art direction here in the way Caves of Qud or Cogmind have art direction. There’s a competent sprite library serving a systems game.

The ASCII mode, oddly, ages better. It’s honest about what it is, it’s information-dense, and it’s arguably easier to read at speed once your brain stops looking for pictures. Being able to toggle between the two is the right call — a newcomer can start with tiles and graduate without restarting anything.

Performance is a non-issue by design. This is a turn-based game rendering a grid — a 2D game of this scope is light on hardware. The technical concerns worth raising aren’t about frame rate — they’re about bugs, which is a more persistent problem in a game with this many interacting systems. A game with tens of thousands of items and hundreds of monsters will have edge cases, and this one does.

Steam is the correct and effectively only home for this, on Windows, macOS, and Linux — a three-platform release from a small team, which reflects the game’s Unix-adjacent heritage more than any modern porting strategy.

Mouse and keyboard, emphatically keyboard. The Steam version supports mouse input, but playing ADOM with a mouse is like using a spreadsheet without formulas — technically possible, structurally wrong. The game’s command set was designed for single-keystroke access, and fluency means learning the keys. Anyone expecting a controller-friendly experience should look elsewhere; this is a desk game.

The platform-specific additions earn their keep. Weekly seeded Challenge competitions turn a solitary game into a shared one without touching the core design. Ghosts from other players’ deaths add ambient community texture — you find where someone else fell, and it means something. Achievements and trading cards are table stakes, but 70+ achievements in a game this dense function as a soft quest log, pointing at content you’d otherwise never find. Cloud saves and the Steam overlay matter more than usual here, because the overlay is where you’ll keep the wiki.

ADOM is a monument, and monuments are not comfortable places to live.

What Biskup and his collaborators have built over thirty years is the deepest expression of the classical roguelike available in a commercially supported form — a game where the corruption system alone justifies study by anyone designing risk/reward mechanics, and where the sheer volume of interlocking content means hour fifty still holds surprises. The Steam release did the right work: three modes for three temperaments, a tileset that opens the door, weekly seeds and ghosts that build community without compromising the design, and a no-DLC promise most of the industry can’t match.

It also refuses, fundamentally, to meet players halfway. The interface is a 1994 artifact with 2015 polish. The difficulty is opaque rather than merely hard, and the community’s answer — read the wiki — is an admission that the game doesn’t fully teach itself. Updates come slowly. Bugs linger. For anyone with a normal tolerance for friction, this will feel like homework.

We rate it a Good: a solid, remarkable-depth game whose flaws matter enough to limit who should buy it. Buy it if depth is what you’re actually after, and go in knowing the first ten hours are tuition. Everyone else should keep walking.