Monday, March 12, 2018

Gygax on Dungeon Design

I've always struggled with dungeon design. Nowadays I finally do have a process which gives me some amount of traction, which is like (1) map out a dungeon (likely with the GridMapper application), (2) outline the monsters & treasures in each area with a spreadsheet, and (3) write the descriptive text for each area with a word processor. Nothing earth-shattering; but it did take me a few decades to settle on this. (Also, the spreadsheet step gives me a convenient opportunity to crunch statistics on the overall dungeon.)

Coming back from GaryCon, it's on my mind that many of us at this point are aware that Gygax himself played with very minimalistic dungeon keys -- as evidenced by inspection of the famous photo above, which appears to show a listing on the left with just a single line per encounter area, likely only documenting the monster & treasure in each room. Also: When I had a few minutes to chat with Ernie Gygax on Saturday, he brought this same thing up. Coincidentally, I recently came back across a fairly lengthy answer on the old ENWorld Q&A thread in which Gygax Sr. explicates this, and expands on the fact that all other setting/dressing information was made up improvisationally during the play session.

I think this is worth highlighting (particularly for new DMs and players) for a few reasons. One: This is not suggested anywhere in the published rulebooks, and is in stark contrast to the published dungeons with the company imprint. For example, the 3-room sample key in the DMG is very wordy (and I've previously looked askance at module B1 for likewise presenting a model of very complicated room descriptions). Two: It explains the difficulty he had in converting his Castle Greyhawk to a published product, because the style of his play notes, and the style he expected for a published product, were worlds apart (as he describes below). Three: Perhaps this gives additional credence to the one-page dungeon format which I normally resist -- or at least a two-page dungeon format, perhaps.

In retrospect, I really, really wish that this process had been explicated in the original dungeon-design advice in the core OD&D/AD&D rulebooks -- not effectively kept masked/hidden. It's among the most important revelations that could have been included, I think. Here's the full answer from the ENWorld Gygax Q&A, dated 15th September, 2005, in response to a question from user Clangador:
As to your questions, I usually made one-line notes for my duneon encounters, from around 20 to 25 of same for a typical level done on four-lines-to-the inch graph paper-a few more on five-, six, or seldom used 8-line graph paper, the other spaces were empty save for perhaps a few traps or transporter areas and the like.

I did indeed create details for the PC party on the spot, adding whatever seemed appropriate, and as Rob played and learned from me, he did the same, and when we were actively co-DMing we could often create some really exciting material on the spot, if you will.

When the encounter was elimiated I simply drew a line through it, and the place was empty for the foreseeable future. I'd give Rob the details of any session he was not at and vice versa, so we winged all of it. Sometimes a map change and encouunter kkey note of something special in nature was made, but not often. We both remembered things well, Rob very well and when necessary something was made up out of whole cloth for the sake of continuity of adventuring.

When new maps were made it was often nearly impossible to have the stairs and other connections line up with other maps, so a note or two and "fudging" served perfectly well, this was particularly true of the means of entering and exiting lower levels from secret locations surrounding the castle ruins.

Now you understand why the Castle Zagyg project is such a major design undertaking. If we handed over the binders containing the maps and the notes don't think even thge ablest of DMs would feel empowered to direct adventures using the materials. ..unless that worthy was someone who had spent many hours playing with Rob and me as DM.

I have laid out a new schematic of castle and dungeo levels based on both my original design of 13 levels plus sideadjuncts, and the "New Greyhawk Castle" that resulted when Rob and I combihned our efforts and added a lot no new level too. From that Rob will draft the level plans for the newest version of the work. Meantime, I am collecting all the most salient feature, encounters, tricks, traps, etc. for inclusion on the various levels.

So the end result will be what is essentially the best of our old work in a coherent presentation usable by all DMs, the material having all the known and yet to be discussed features of the original work that are outstanding. .I hope :uhoh:



  1. I've never had a campaign be enjoyable because of what is planned; it's always enjoyable because of what was spontaneous. Plan as minimally as you can because the several players (yourself included) will bring the fun regardless.

  2. My hats off to the GM's who can run published adventures direct from the books.
    I can't keep track of multiple players and npc actions and lengthy adventure into, so even basic adventures end up getting re-written in my format.
    Page 1 big map, probably with hand written reminders of important things (FLOOR TRAP, ALARM BELL) pinned to the inside of the GM Screen. The rest of the pages set for quick room details (since I am terrible at remembering to put atmosphere in, that chart in the 5e DMG is a life saver). Monsters stats are one line with a reference to the MM page if they have crazy abilities. Then enough space in between to make combat notes.

    It is a funny contrast to the way Gary seemingly ran a dungeon vs what was published in those days. Those old modules were dense.

    1. I've definitely evolved. I used to put a lot of prep work in to be able to run modules by the book (honor the playwright's words, as it were). What you describe in your screen is pretty much the same as I have now.

      And you're right that the disconnect between play and publication has been a lifelong conundrum for me (esp., someone growing up in an isolated rural place who had to learn purely from the written books).

  3. The funny thing is that the OSR seemed to rediscover this way of doing things through its "one-page dungeon" concept. I'm not sure whether the re-discoverers were riffing off of stuff like that ENWorld Q&A, memories of some of the early dungeons, early texts such as Palace of the Vampire Queen or Arneson's First Fantasy Campaign or whether it was, as it were, a true independent rediscovery.

  4. The one part of the AD&D that seems to generate a minimally keyed dungeon in the manner of what Gary actual ran with is Appendix A: Random Dungeon Generation. And of course it is a revision of the Solo Dungeon Generation article from Strategic Review #1.

  5. Coming into DnD last year as a newb with background in improvisational music here’s a perspective: working through published adventures is a lot like learning iconic improvised solos. The exercise helps you learn the underpinnings and language of the medium. You aren’t going to learn to improvise at a high level without some background and base knowledge to work with. Once you have a strong command of the underlying pieces you can break them apart and summon and command them at will deftly. But if you are coming in without being steeped in the experience another way, seeing how others before you have done it is a powerful learning tool. So as I work through Mines of Phandelver I am looking for places I can expand and tailor it to my group but the basic scales and chords assembled into a full lead shear in this particular way give me a robust framework to start with as I earn and learn my own style.

    Cheers, Great blog

    1. It's an interesting parallel, thanks for that! However, I wish that the "real documentation" that the masters worked with had been shared with us (so, e.g., I didn't lose 30 years chasing after overly verbose notes). It would have been handy to see a "minimal dungeon key" next to an elaborated "example of play" to see things getting fleshed out improvisationally.