LESSONS FROM HELL: John Romero's Level Design Rules for DOOM


In December, I finished Doom Guy: Life in First Person by John Romero. I highly recommend the book in large part because it captures and conveys what its like to be caught up in one's own passions. I feel this way about Game Wizards & The Elusive Shift. 

The connection between DOOM and D&D is pretty strong. In fact Romero's id crew was regularly playing Carmack's D&D campaign on Saturdays after their breakneck coding sessions. And in one session (I think the last they ever played) it was Romero himself who cut a deal with a demon to gain a powerful magic sword in exchange for a tome that would give the demons knowledge on how to invade the material plane-- and they did- ending the world. End of D&D and the beginning of DOOM.

In Doom Guy, Romero outlines his level design rules. And I wanted to discuss how they might inform dungeon design. These rules work in companion to Goblin Punch's Dungeon Checklist. Where Punch's list is about what is in the dungeon, Romero's list gives one an idea about how the dungeon is constructed at a high level.

1 | The start of the level should present interesting choices or look impressive

  • Yup! When I run small dungeons off the top of my head, in the first room I usually have a weird statue, a locked door east, and an open arch west with something that can be seen.
  • Hard to make something "look" impressive in our pen & paper context here so instead I would say describe something fantastical about the adventure location that thematically sets the tone for the adventure.
  • I would also say that "interesting choice" or "interesting look" is a 1:1 exchange, so instead of going for impressive vistas try interactable vistas.

2 | The start of the level should fit its purpose. Do I want to teach the player or make them feel scared? If the former, there are no enemies. If the latter, watch out.

3 | Reuse areas in the level as much as possible as it reinforces the understanding of the space every time the player goes through the area again. For example, if players come back to a central halo before going out to a spoke, they will remember the hub the most.

  • I tried to do this with my "wine dungeon" and I think it works quite well because players develop a strong sense of the space. What to avoid and where shortcuts are. 
  • I think players often dislike dungeons because they are rarely allowed to get to know a dungeon and then employ that knowledge to their advantage. 
  • This in turn has resulted in dungeons becoming truncated and further withering of this opportunity. Nightwick has been something I've mapped for the past two years and it really has some important knock-on effects in terms of player planning

4 | Provide contrast in every element of the design: light, sound, and action. This keeps a level fun and interesting and prevents it from falling into a monotonous loop of gameplay. We want the player to feel like they are on an exciting roller-coaster ride.

  • I read this as its better to maintain an aesthetic consistency (variety) than a realistic fidelity (monotony). Meaning it is less important that a tomb is true to its IRL counterpart and more that it has a variety of things that are tomb-like. Or fit with what has been signaled by following #1 & #2.
  • Although not stated, enemies are also party of level design and variety is important there too, but again it has to be within the constraints of what has been signaled already to allow the players to maximize agency through preparation and reasoned decision-making

5 | Changes in wall or floor texture should be accompanied by a height change or border texture

  • I admit this one might be hard to conceptualize in an RPG format, but perhaps its best interpreted as provide transitions between two different aesthetic areas.
  • If you have a burning hell pit next to a golden throne room, perhaps its best to have a room, hallway, or corridor statues with melted gold crowns and scepters and scorched tapestries.

6 | Include at least four secrets in your level

  • This feels like one of the most D&D pieces of advice and still just a really solid one. Why not include 4 secrets in your level? Honestly, I bet this is also a single trick that would make most D&D dungeons 100% better.
  • Also, this doesn't just have to be singular doors, both early D&D and DOOM would hide secret doors behind secret doors.
  • Often DOOM would also provide an audio cue that something has happened, but its not always immediately obvious what.
  • DOOM also is good about showing a goal (through windows in a wall or door on a floor of lava), but not how to get there.

7 | When a player solves a piece of a puzzle, they should already know where to go next. An example would be that you already tried to open the red door before you found the red keycard. A bad design would be to flip a switch, then see and hear nothing that shows you what you just did.

  • I think the ur-concept here, on a very high level, is that when players affect change in the environment, they should see the effect of that change and understand how they can go further with it (or if they can't, they know this as well).
  • The second, more specific, point is good too: show the problem before the solution. Or you risk players not being able to recognize the solution. This seems to be an issue more with mystery games, but puzzles and locks still show up enough in D&D its still important.

8 | If an area in your level looks like it could be made in an earlier tech, you have failed. Make the area more interesting and use more of the engine's features to ensure that.

  • Ha! Might be damning advice for old-school play considering the scene is concerned with using "old" D&D tech. Also kinda funny in light of Romero himself putting out new levels for DOOM (1993) in a package called SIGIL and SIGIL II✤ using said DOOM.
  • Here is what I would transmute this advice into:
    If an area in your level looks like it could be made [using only random die rolls] you have failed. Make the area more interesting and use more of the [RPG system's] features to ensure that.
  • I think this kinda harkens back to my "03 Tomb With Eight Skeletons"- don't stop just at random stocking with by-the-book tables- alter something- change the "tech". Generate new encounter tables, new monsters, create a more interesting room feature ect.
  • Use Tomb of Adventure Design or The Monster Overhaul or reskin to a more exciting effect 

SUMMARY: I don't think that all of the above 8 points are unknown to those who swim in the classic-play dungeon exploration sphere, but they are concise and pithy. I think they are worth considering and might be useful to reflect on when trying to decide if your recent dungeon work or effort is all that it can be.

✤ I've also been playing through SIGIL and SIGIL II. What a delight! DOOM (1993) is to FPSs as Moldevey/Cook BX (1981) is to D&D. Highly recommend.


  1. Solid stuff. You can see how the Doom designers were thinking about the same questions as D&D people regarding how to convey narrative through space, yet had a totally different set of tools. Surprising but not at all surprising that most of the guidelines can be retrofitted back onto D&D.

    1. Exactly on a high-level there is a desire to make exploration apart of the game as well as attention to the map.

  2. Thank you for this Warren, interesting perspective - I really like how you have re-interpreted #8 for imagination gaming. Reminds me of Ty's post on Mindstorm about Question-Based Design and Chris McDowall on a podcast (apologies I can't recall who was interviewing him) about 'seasoning' your ideas but adding one or two layers of unexpected details (manifest in interpolating results from spark tables). As a first pass "2d6 goblins" is fine, but I what I really want at the table are one or two unusual spices: "war-painted and wailing as they bear a fallen kinsmen to his final rest" or "clobbering each other over who gets to keep the mugged wizard's tome screaming 'knowledge is power!' " You put it well: certainly start from random tables and generation, but it shouldn't still look like that when you're finished!

    1. Thank you for reading! You put it even better than I did. Much like any writing or creative endeavor, completing a thing is the primary goal. But then having the decipline to put it aside and return for editing passes really makes it sing.

  3. Good stuff. Like the 4 secrets per level checklist item, especially.