Tuesday, January 19, 2010


Ok, so the poison dart room is here, and then the ice-statue room is here.

No, wait, I'll do it the other way around.

No, wait, actually, it doesn't matter does it?

I mean, I'm pretending it does, but really, either way, it's a dungeon and the PCs can go wherever they want and they're just as likely to wander into one as the other. Or both, or one first and then the other.

I mean, I could just have a blank map and roll randomly for every room.

Ok, no, that's not exactly true, some rooms have a teleologic--boss goes at the edge, eerily quiet room goes in a very specific place, the mystery tunnel goes there, the plot-seed rooms must be choreographed with some care, the gears must be next to the grinding room, etc...

But, honestly, half the rooms, they could go anywhere, right? Why pre-determine them? Just have a random table of random rooms like wandering monsters.

Ok, so:


-it's a small dungeon where the PCs won't level up in the middle, and...

-the "plot-important" rooms, and rooms that mechanically relate to each other are already properly placed and properly spaced, and...

-whatever architectural model the dungeon is supposed to represent is accurately reproduced (like if it's an old house it has bathrooms, kitchen, etc, in basically the right place).

Then why not just have an empty map with a random table of pre-written rooms and you roll on it each time the PCs go into a room?

Pros? Cons? Discuss...


  1. Well, it'd save time and the occassional bafflement by doing everything beforehand. Sure, you could roll for the entire dungeon, and the AD&D 1e DMG had those very rules in them. But rolling while game is going on? I dunno, seems time-consuming; the less you have to roll and consult during the session the better (I know, that's blasphemy coming from a Hackmaster player!). And if you roll everything beforehand, that's the same thing as NOT using the random way of doing things (i.e. before the session you populate your dungeon by any means).


  2. Absolutely pro. Actually, I feel that for megadungeons random generation is a must.

    When I ran the old Advanced HeroQuest game, I really enjoyed the random generation of dungeons. In that quests, you only mapped and staffed the last level of the dungeon, the quest level as it was called. And even that level could be randomly created.

    That kind of run-by-the-seat-of-your-pants GMing was incredibly fun for me. I had to think quickly how to describe it, how to place the monsters and all that. Actually, I liked it so much that used the random dungeon tables in my AD&D games (they were all d12 based!).

    The only con I find is the rare chance of getting nonsensical results, but I feel tha can be prevete with a good design.

    By the way, excellent blog :)

  3. Pro of yor proposal: can't see really anything, except the fact that you don't have to plan as much.

    Cons: if you roll the new room just before entering, you might miss the fact that the two rooms might interact. Plus, it's time consuming.

    I'd recommend, if you really want to roll up rooms on the fly, to roll up the new room and the adjacent rooms as well, or to roll everything beforehand and see what kind of emergence is available.

  4. Looking back at your mapping post, I don't see why not.

    I'm thinking if it's a manor house-type thing, as long as you're not randomly putting the master bedroom down next to the kitchen or below a wine/root cellar, it should work.

    I think it depends on just which rooms are "plot-important", like if the party is a band of thieves breaking in a darkened second story window to "remove a statue" somewhere in the house, that first disused room with the window becomes pretty important...

    Are you linking these "p-i" rooms with set hallways and then adding the random rooms to fill in the rest of that building's shell?

    If not, that could be tricky for a dungeon crawl.

    I'd also might use a little table for the # of windows/stairs/doors and maybe a chimney or two big enough to let anybody out of those rooms/houses (and those not big enough...).

    You also seem to have a mapping style that easily works well with the geomorph approach, so I'd say make eight or so differing building shell templates and try filling it with some differing, standardized room shapes.

    Did you ever find any good city source books?

  5. Sham's Empty Room Principle applies. You need space to drop in all the good stuff you improvise during play. ;)

    I suppose you could reformat the OSRIC random dungeon tables into a nested format so that they're quick enough to use in play...

    @Imperator: Good catch. I'd completely forgotten about the random dungeon levels in AHQ. That's probably because the mapped and keyed 'set piece levels' got more play in the books & magazine articles.

    @biopunk: Warhammer City or Lankhmar are my urban go tos.

  6. I like these kind of random methods, they are absolutely right, and they save time too. It even feels more real to me, city or dungeon or wilderness, if there are lots of random elements in it. I like to make my random rolls before play, so they can give me even more ideas of whats going on in the place, but it can work during play too, and then I'm exploring with the players.

  7. Since I often get my best stuff by rolling randomly and being forced to do something with it on the spot, I have strongly considered going with the on-the-fly random method for all but a small number of key areas of the dungeon.

  8. Or forget rolling entirely & put whatever room you FEEL like-- like if the players are feeling freaked out & on edge, put the Darkness & Ambush by scuttling monsters room. Or if they are bored put the "mop the floor with these guys!" room. Or whatever.

    Me, I don't like it. I like to build the area, them scatter interlinkages throughout. Glowy crystal cave? Okay, so the goblins have glowy crystals instead of torches. Ancient tomb over here? Okay, so these ape-men have ripped down the friezes to decorate their cave. Etc.

  9. Random generation is great for HUGE areas (the megadungeon example above) but I find it less useful for smaller ones (the manor house). The tables can help you keep an area fresh (remove a room from the list once you've used it) or "living".

    A seperate page for each encounter that you plan to use and then remove from the binder/notebook/etc would be my preferred method of organization (you could reshuffle/renumber them as you progress onto different adventures).

  10. I saw a variant on this in a d20/ogl game where each time you rolled in the "threat range" you encounterd the next key-plot room, and each time you didn't the treat range increased by 1.

  11. As you stated, mechanically, what's the difference to the game/players? Nothing, of course. I think this lies solely in the realm of DM comfort.

    I could not operate in this way. Sure, plenty of off the cuff things happen all the time in my campaign, but the rooms in a set building are always set up prior. I like to look to see what the next 3 rooms are while the players are making decisions/noise so creatures can stream in appropriately.

    Our game could be very different from other's games, though, as the combats are really tactical, terrain oriented affairs.

  12. This is one of the key differences between old school and new school adventure design.

    I remember giving the following example to my friend Yuri, a fan of the complexity and randomness that is Rolemaster, many years back...

    "You are searching for the Eye of Zorm. Before you can even locate it in the dungeon, you need a cloak called Zorm's Lid or the Eye will fry you the moment you come within visual range."

    I then describe a few halls in the dungeon where the Eye is kept. Finally there is the traditional fork and I describe both hallways in detail. I then ask the obvious question, "Do you go left or right". He decides to go left.

    A few more winding paths and example encounters later and he reaches the Eye. The End. This was not an actual game you see, just a 'for instance', hypothetical example. I than asked Yuri, what did he think would have happened if he had gone right?

    "I wouldn't have found the Eye", he said.

    My answer was, "No, the encounters before reaching it might have been different but if your character survived you would definitely found the Eye. The Eye is the point of the adventure. Its going to be there at the end no matter where the end is."

    To me the 'dungeon' is scenery and setting and certainly an obstacle but it isn't an obstacle for me the GM to tell a cool and exciting tale of adventure. I build my 'dungeons' organically in a manner similar to what Mordicai is suggesting. Largely, the interior is ad libbed with only the most generic map. Usually I'll take an unlabelled map from a book or the internet and I don't even have anything written down for what's in the rooms. Instead, I have a story, a theme and the vaguely logical progression of stuff that fits the plot.

    20'x 30' rooms full of nothing just don't do much for me.

  13. Well, if it's a house or something, there's usually a logic to the rooms - like the kitchen and the larder and the dining room will be close, the bedrooms are usually on the second floor, etc. I don't know - as much as I like random charts for stuff, I prefer a keyed map with only a handful of empty rooms for on the fly stuff. But then, I REALLY enjoy drawing maps and designing dungeon rooms - it's one of the reasons I love to DM - and I put a lot of care into it. Your method is probably faster and still fun but I prefer to place my rooms strategically, to slowly build the exact atmosphere I want from room to room. Just my 2 cents.

  14. Do the rooms have to make sense? Even using the manor house example, if its a ruin and been unoccupied for a long time, the rooms may no longer betray their original purpose.

    I am also very pro random rolling-as you said what does it matter to the players? Apart from some plot specific opponents, I believe those random encounter charts are engines of imagination.

  15. Very much pro in my opinion. I have a long list of 'random encounter rooms' - just rooms that could show up anywhere in the current dungeon. Keeps the numbered key smaller (there's a few anchored rooms), but for the most part, the rooms are unkeyed and there's a 1 in 6 chance that a room will be one of the 'special list' rooms

  16. My answer was, "No, the encounters before reaching it might have been different but if your character survived you would definitely found the Eye. The Eye is the point of the adventure. Its going to be there at the end no matter where the end is."
    This is sort of how I think I'd like to go about scenario construction, but I'm always wary to do it, because I worry about player agency. If whatever their choices, they still get to the Eye, then haven't I then robbed them of choice? Haven't I then railroaded them, even if they don't know it?

    There's an invisible (or visible, if you have a GM screen!) line between the GM and players, and until they cross it, there is no difference, to them, between choice and the illusion of choice. But something about it seems dishonest to me, so I can never go through with it.

    So yeah, I'm in favour of the spontaneous dungeon in theory, but I don't know if I could ever run it without second-guessing myself.

  17. There's actually a historical precedent for this kind of dungeon. When Ken St. Andre played his newly designed Tunnels & Trolls with his friends in the Phoenix Cosmic Circle, his dungeon "Gristlegrim" was a set of index cards where he picked/randomized a dungeon room after the other. Tried and true, in other words. It's a method of the seventies! :)

  18. Wandering room rolls every turn?

  19. In practical terms, GrimJesta's point about time rings true to me. It's a mixed bag playing it this way -- some advantages, some disadvantages.

    I usually run D&D with random stocking on-the-spot, for a couple reasons. One is that I never get around to doing much prep. The other is that I'm never really all that sure what target the players in my "heist-style" urban D&D campaign are going to rob, anyway, so a lot of prep would be wasted anyway. I just have a bunch of mansion floor plans and dungeons for basements, and I roll to stock rooms as we go. (using the dungeon stocking table from p. B52 of the Moldvay Basic Set).

    I will say I've gotten some damned fun results from it, which is something it has in its favor. But there's no question that it's also kind of clumsy in play, in my experience. Not clumsy enough for me to actually replace it with traditional prep, but it is kind of clumsy. Also, like tsojcanth said, it makes it more difficult to play up the interactions between nearby rooms. And it also means that you can't give players information about their environment that would allow them to make non-random decisions about which direction to go. ("It smells of sulphur down the lefthand path." Etc.)

    OTOH, your version of this might work better than mine. It sounds like you'd be prepping and creating the tables yourself. This would mean you'd have a much better feel for what's in the tables, and you'd be able to arrange them in an intuitive way for use in play.

  20. Nethack is still the next best thing to actual D&D.

    Which is to say, yay procedurally generated content.

  21. pros and contras

    Hey, Adam.