Sunday, February 2, 2014

Megadungeon Conundrums - Pacing Versus Size

Gus L runs interesting megadungeon campaigns over at Dungeon of Signs - it's a creative, fun place to drop in.  Today he's wondering about empty room theory, the pace of exploration, and dungeon size.  (He's referencing Anomalous Subsurface Environment, Pat's published megadungeon, but the discussion applies to pretty much any megadungeon).

Empty rooms are important.  They support verisimilitude - the dungeon isn't a packed zoo, straining credulity by stacking antagonists next to each other in the dungeon.  Empty rooms support space for lairs and buffer zones and areas for wandering monsters.  Empty spaces reinforce the theme of exploration and build tension into the game.  But when the sessions are short, folks don't want to spend all their time in empty rooms, either.  The larger the dungeon, the more empty rooms, and the more burden on the referee on making that "between time" interesting or motoring to the next set piece encounter.

Our approach has to be more nuanced and account for more variables - the size of the level, how much do you expect the players to explore before descending, the number of players, and the expected fatality rate.  Players tend to be completionists, unless there's a strong impetus to descend.  A party may level up before "completing" the first level, such that the second level isn't a challenge for them by the time they get down there - they'll be well on the way to their 3rd level of experience.

I tend to take the average party size (for my group, that's 6-7 players), and assume they'll level up when they get to 15,000 or so party experience points.  Some XP will get drained by henchmen, some will vanish when the first level characters die.  So maybe the whole dungeon level needs about 20,000 xp distributed to account for leakage.  The blend of monster vs treasure XP varies by edition, but it's usually 80-20 or thereabouts (80-90% percent comes from treasure) - so a first level dungeon could have 15,000 - 16,000 in treasure XP (gold or silver standard, depending on your game).

If there's a strong impetus for the characters to move down before finishing the level, make an assumption on how much you expect them to complete - 50%?  Than the total wealth of the level should double!  Once they clear half of the rooms, they'll be ready to descend.  If you expect them to grind through every last room, than distribute the base amount evenly.  You can see that a 100 room first level is going to feel much differently if each room averages 150 xp of treasure versus a 25 room dungeon where the room average is 600 xp of treasure.  Room density is a factor as well - empty rooms versus monsters\traps\specials - as well as how bland are the empty rooms versus the level of interest in the empty rooms.  It's a lot to consider - no one said dungeon design was easy!

But Beedo - isn't dungeon design a mystical art, irreducible to mere mathematics?  What is all this number-crunching and discussion of average XP per room have to do with My Homebrew's (TM) amazing flavor?  Get your math out of my fantasy.

Running a long term megadungeon is no picnic.  I see groups flame out all the time here in the blogosphere, or declare megadungeons as boring, and the biggest culprit seems to be pacing and the grind of exploration.  When I was in 8th grade and we played all day long on the weekends, those 10 hour epic game sessions fueled by chips and pizza, we had the time to take a 'completionist' approach to exploration and were willing to collect coppers and salvage dented helmets and whatnot to squeeze every last experience point from a place.  Nowadays I'm lucky to get the adult gamers together for 3 game sessions a month, 3 hour game slots.  My people don't have patience for endless dungeon levels where they're expected to clear every last room.  They want the experience of being in a vast place, too big to fully map, but don't actually want to visit every room, themselves.  (It's enough to know it's out there.)

All is not lost.  It is the referee's job to pay attention to pacing and keep their hands on the dials.  If the players are doing well XP-wise, and ready to go down a level, have some things happen that make the rest of the first level unattractive, or provide a strong incentive to skip the rest.  Perhaps rival adventurers clear out many of the remaining areas when the party is back in town, so that the second level is the next "virgin area" to explore.  I've had patrons and similar figures provide missions, quests, and related hooks that keep the action moving down instead of horizontal, before the challenge level wears off.  If it's your own dungeon, make the levels a little smaller and modular, so if the party is made of "completionists", they can still finish the whole map but descend quicker to a new place before any sense of grind crops up.

For the Black City campaign, the dungeon level was vast and sprawling, but it assumed players would level up as they went further from the entrance.  The first level wasn't truly for level 1 guys - it covered levels 1-3, and only the nearby zones were for level 1 guys.  In that one, I did have rivals clear out the other near-areas, encouraging the players to push deeper to keep pace with their growing power level.  Likewise, the second level is really covering character levels 3-5.  Of course, if your dungeon is truly "mega", you have to be willing to let some of your work product go unused in the interests of pacing - unless your players really want to see or map every last room.  In which case, according the math model, you need to draw down the average XP per room so they don't blow the level caps before making it down.

For Harrow Home, the new project I started here, the plan is to have smaller levels, but have lots of them, so the forward impetus is to go deeper into the dungeon.  It's a vertical structure rather than horizontal.  It sounds good in theory.  Will it feel like a megadungeon?  We'll see how it goes.

I'm a big fan of works like ASE and Stonehell.  I do believe megadungeons with those sprawling levels and a hundred rooms can work well.  However, I would definitely take my math-based approach to determine the relative wealth of the levels, identify a fair percentage the players would be expected to complete, and then put some campaign events or drivers in the game that would encourage them to descend before the game reached any level of grind or repetition.  If the math indicates the players need to grind a hundred rooms before advancing, make sure that fits the group culture and play time, or increase the treasure amounts and reduce how much they have to explore before going to the next level.

If you find this type of stuff interesting, I did an analysis of some published adventures a few years ago when these ideas first start percolating around my skull - We Came, We Saw, We Leveled Up, and Treasure By Adventure Module.  They break down things like expected levels earned, amount of treasure, and density, so you can draw your own conclusions about which adventures support the Monty Haul experience versus forcing your players to grind out the coppers for a chance to move ahead.


  1. I think a distinction made between truly empty and merely unoccupied would be helpful. I am not sure which is being discussed in most cases.

    I sit on the players side of the screen so infrequently (it has been at least 5 years at this point) that I probably don't have any useful data to contribute in that regard. However, leaving a truly empty room (as opposed to merely unoccupied) makes me feel lazy (and bored) as a designer.

  2. I prefer to go with unused or unoccupied rooms as well - the room has\had a purpose, and that lends some direction to the details.

    I like the idea of Gus L's table of things for empty rooms though - from the post lined above - it's a bit of time-saving quantum fun. I can see myself making future dungeons that are so old and unknowable that the original purpose is long forgotten, and planting something interesting in there (how it's been used recently) would be more appropriate.

  3. I started taking the math approach to my campaign's megadungeon, but I didn't take it fully. I actually have been writing the thing on a spread sheet, and code each room as empty/monster/monster+treasure/, etc, and tally the result. I had empties (which I still strove to make interesting) at 30%. While I also could and did randomly assign and then design the contents, I manipulated them to create areas and factions, and kept track of my 'progress' on the percentages.

    I had separate columns for base description (usually static fixtures), contents, monsters, treasure. I didn't follow through well enough to keep columns separate and haven't cleaned it up to tally exp per level.

    Lately I've been wondering at restocking questions. Like with many big projects, the difference between intention and execution has been interesting and there are things I'd do different if I had to start over. But there's enough progress that I would shudder to have to.

  4. I'd say, write in rooms regardless of whether the players will see them. The cooler you think something is, the closer to the core of the dungeon you should stick it, and the nearer it should be to passages up or down. That will make it more likely they will see it.

    Second, divide incoming XP using a fraction of (Dungeon Level) / (PC Level). That is, 1st level PCs on the first dungeon level get 1/1 XP. If they reach 2nd level and they continue to adventure on the 1st dungeon level, they get 1/2. But the maximum is 1/1, meaning 1st level PCs on the 6th dungeon level don't get 6/1 (or 6x XP awards), just 1/1 (1x XP award). The rule is thus exclusively a penalty for tarrying where you kick too much butt and learn too little.

    The XP penalty should be for monsters and treasure too.

    So consider a 2nd level party deciding whether to descend to Floor 2 or continue clearing out Floor 1. If they stay on Floor 1 they'll probably get loot etc. but it's Level 1 loot, and they get half XP for it. But if they descent to Floor 2 they're getting Floor 2 loot and full value for it too! That's the sweet spot. They could push their luck and descend to Floor 3, with commensurate risk and reward, but it's just standard Floor 3 reward rather than multiplied by some amount.

    I believe something like this was in 0D&D and also in the 1E DMG.