Coelacanth: Lessons from Doom | vector poem

This post is intended as a companion piece for the release of Arcadia Demade.

A high-minded goal like “expand the boundaries of the medium” doesn’t always mean forging ahead in crazy new unknown directions. Sometimes it means examining lost evolutionary lines in game design – picking up ideas that were abandoned long ago and seeing if there’s any new life in them. The game I keep coming back to in this regard is Doom. Not the 2004 reboot, but “Classic Doom”: Doom 1 and 2, Final Doom, the Master Levels – and its vast universe of user-made content. What can it teach us today?



In 1993, the message Doom sent to the videogame world was something like “use cutting edge technology to make something dark, edgy and violent”. The world has changed so much around Doom since then that very little of that original impact comes through to players today – though the industry has inarguably gone on to master the techno-fueled ultra-violence thing! Here’s what I’ve found after many years of enjoying the game and digging ever deeper into its design:

Doom feels more like 1st person Robotron than a modern FPS

When you play Doom today, it doesn’t feel much like you’re controlling a human or moving through real spaces. Try this though: press the TAB key, type IDDT twice and pretend you’re playing Geometry Wars, and the moving triangles are your enemies. This is what Doom’s designers were working from in 1993 – back then, the idea of a first person shooter was barely established, and their closest models for many mechanics were from 2D shooters like Robotron, Berserk and Tempest. This approach echoes throughout Doom’s design. The notion of realism in FPS design wouldn’t appear for another few years, and many decisions were made simply on the basis of being good for abstract shooter gameplay.

Partly thanks to this, many parts of Doom’s “game feel” still compare favorably with modern twitch games. Enemy speeds and patterns are very finely tuned, weapon design is strongly orthogonal, player movement has a nice friction to it and level design elucidates all of this. Quake 3 is still considered the pinnacle of arcadey FPS movement and feel, and that lineage starts with Doom – some of the code is even similar.

Doom is about “maneuverability as defense”

In almost every modern FPS, the player moves fairly slowly and a huge proportion of enemies are equipped with instant hit attacks – pistols, machine guns, sniper rifles. This usually puts the player in the role of “damage sponge” – they’re intended to soak up a certain amount of damage from mostly unavoidable enemy attacks, then seek cover and heal up. Halo’s recharging shield makes this mechanic quite explicit – by default, you’re exposed to damage and will die, while seeking cover halts that and completes the basic cycle of any combat.

Contrast all this with Doom Guy, who runs at about 50 scale miles per hour – nonsensically fast by modern standards. Most of Doom’s enemies don’t have instant-hit projectile attacks, and most of the ones that do are quite weak – the lowly trooper and sergeant. Every other enemy projectile takes time to reach its target, and would look comical in a more realistic visual presentation.

So because the player moves so quickly in Doom, and because most enemy attacks are dodgeable, the player can avoid a significant amount of damage simply by moving. A skilled player can often deal with large numbers of enemies sustaining hardly a scratch. This creates a feeling that’s quite rare in modern FPS: that you are powerful because you are agile, not because you’re a tank. This frees up Doom’s encounters to feature huge numbers of enemies, to vary scenarios by mixing in different proportions of threats, and to have huge, sprawling, often non-linear spaces that the player can traverse easily. There’s nothing quite like it today.

Doom has a more varied bestiary than most modern FPSes

In many modern FPSes, the design of every enemy the player faces is sampled from a fairly narrow tactical spectrum – soldier with machine gun, soldier with shotgun, zombie with melee attack. Doom, on the other hand, has a huge range of monster sizes, speeds, strengths and movement/attack patterns. Former humans and imps are slow moving ranged fodder. Hell Barons are large, tank-like threats. Flying enemies range from the small charging Lost Soul to the tough, fireball-belching Cacodemon. Revenants and Mancubi launch homing and spread-fire projectiles respectively, and the three boss-class monsters are each very dangerous in different ways. Some enemies can be stunned by weapon fire more easily than others.

Such diversity creates a large but simple to understand toolset that level design can combine with architecture to create a huge variety of combat setups. One tough guy with a lot of fodder means the player has to do crowd control while focusing on the real threat. Lots of flying enemies make the player seek low cover and choke points. Enemies with strong melee in tight spaces make the player dance and really exploit the stun properties of their weapons. This versatility of the core design makes life easier and more fun for the level designer, and thus the player.

Doom was abstract in ways that empowered its level design

While some of Doom’s levels have a very thin fiction via their title (eg “Hangar”) and general texturing theme, if you actually explore them you find they only resemble real locations in the loosest sense possible. This is precisely what allowed Doom’s level design to present a wide variety of interesting tactical setups. Level designers didn’t have to worry about whether a change made something look less like a hangar or a barracks, just whether it was better for gameplay. This was especially critical for a style of game that was just finding its feet in 1993.

As the march of technology has allowed ever-higher graphical fidelity, virtually every FPS since Doom has attempted greater and greater representationalism with its environments. While games like System Shock began to show that a real sense of place can be a huge draw in itself, designers of such games will always have to manage the tension between compelling fiction and optimal function, unless you are willing to go all out and have the kind of weird, abstract spaces Doom has. I would love to see more modern games break with this conventional wisdom and see where it leads, if only in an indie or experimental context.

Doom enabled a revolution in player-generated content

Though advanced for its day, Doom’s technology was still simple enough, and its content low-fidelity enough, that a huge mod community coalesced around it to produce an unparalleled number of levels, mods, total conversions and other addons. This, combined with the fact that the player base was so focused on a single game, means we’ll probably never see something like it again. The lesson for future games might be this: make your technology extremely simple, easy to modify, ship it with a diverse enough pool of content that people can extend it to create a variety of settings and styles, and promote the sharing of this content as a way to add value to your game.

Many PC games have gotten all that right but failed to attract a huge community because of the content fidelity issue. The barriers to entry facing someone who wants to make a mod for Unreal Tournament 3 today are vastly higher than those facing a Doom modder. You can rough out a Doom map in a few hours and finish it in a few days, while that same amount of time might produce a single texture for a modern game. Again, this is something we could branch out from if we lose our fixation on technology and high fidelity visuals uber alles.

Another unique side effect of Doom’s simplicity is that its design principles can be synthesized and expressed procedurally. Level generators for more modern games have been attempted and abandoned, while the Oblige random level generator creates a decent Doom level with proper combat and resource balance, key gating and architectural themes.

Doom is one of many classics whose less obvious qualities are seldom revisited

Doom’s impact has faded, and its precise recipe for success is unlikely to be replicated; nevertheless, the game industry has become quite adept at mimicking its superficial qualities. However we as creators and critics owe it to ourselves to look at Doom, and other classics of comparable depth – M.U.L.E., Ultima IV and Star Control II are a few examples I would offer – and trace less-traveled paths of analysis in search of deeper truths.

Sometimes we must look to the past for guidance. Other times we must strive to forget it entirely. In the balance of both, we will find much to learn about making the games of tomorrow.

Addendum: Not sure if he even remembers it, but Nathan McKenzie made some observations on Doom about 7 years ago now(!) that set my wheels turning on this post, so I’d like to thank him for those initial insights.