Friday, January 2, 2015

D&D 5E and the Megadungeon

I've been enjoying the 5E rules.  It looks like a good version of the game.  But I keep coming back to how I'd run my style of campaign (preferably involving a large dungeon, nay, dare I say it? - a megadungeon) using the 5E rules.  Here's the problem:  characters level up extremely quickly in the first few 5E games.  You don't need a lot of game content before they're ready to move on to bigger and better things.

To put it in perspective, it's common for an old school megadungeon to have 80-100 rooms per level.  About a third of those rooms will have monsters, and players should be in a position to level up after 10-15 encounters in an old school game - about half of the rooms.  I've run a number of megadungeons the past few years.  It usually takes anywhere from 4-8 game sessions to advance, depending how long you play.  I could go into an explanation how I came by those numbers, but it's not terribly germane here - I want to focus on 5E.

The 5E characters don't spend a lot of time in the early levels - they should level up from level 1 to level 2 after the first game session, and get to level 3 after the second.  From there, the experience curve increases such that it should take 2-3 sessions per additional level.  It's still a fast ascent, about twice as fast as the older editions.  I don't disagree with the approach.  I can see younger gamers appreciating the chance to constantly improve their characters.  For the referee, there's no chance of getting bored with the bestiary - you're constantly throwing new and different monsters at the players.

But let's get back to that problem regarding the content.  You don't want to invest a lot of time creating content for a large dungeon, content that has very little chance of seeing play, because the players are advancing so quickly.  So those are our constraints - fast advancement, and limited prep time.  Constraints breed creativity.  The outside the box approach is to just change the rules.  Nuke the default experience system, and start using gold for experience like an old school game.  Make it work just 1E AD&D.  That's easy, but it's kind of lazy.  Can't we do better?

First, I tend to think of two styles of megadungeon, closed and open.  A closed dungeon is lost or unknown to the world, and the players are the first people to explore it after a very long time.  A closed dungeon is like a ship in a bottle or a fly trapped in amber - if there are living inhabitants in the dungeon, it must be some kind of closed loop balanced ecosystem in there, at least until the players show up and start killing everything in sight.  Designing a closed dungeon to fit the players seems pretty straightforward - just size the levels properly for advancement.

The other style of dungeon is different.  An open dungeon is well known, and visited.  It's assumed areas of the dungeon have been mapped and explored, perhaps cleared out by other adventurers.  There's foot traffic involving rival parties, and monsters from the surrounding areas may come and go.  It's a much more dynamic environment.  I tend to prefer open dungeons, it's fun to think through what it means to the surrounding areas to have a known source of adventure, danger, and dungeon gold just sitting in the nearby wilds.

Here are some options that are going across my brain for a large dungeon that attempt to work within the constraints - fast advancement and limited content.

The upper levels could be small.  Perhaps they are geographically small (meaning the maps don't cover a lot of ground), or the maps are large and relatively empty, and the content coverage is small - like maybe there are only a few areas of the upper levels that are unexplored territory.  A small map works fine for a closed dungeon, the large empty map (with only a few unexplored areas left at the fringes) is fine for an open dungeon - especially if the players can learn about the unexplored areas in advance, through rumors back in town.

What about collapsing levels 1-3 down to a single large level?  5E was supposed to bring 'bounded accuracy' to the table, so the danger curve isn't as steep.  It could satisfy the old school blood lust for a high mortality rate amongst freshly minted characters by mixing in 2nd and 3rd level threats throughout the dungeons upper works.  Dungeon level 1 would cover character levels 1-3, dungeon level 2 would correspond to character levels 4 (and possibly more), and then each new level would have a corresponding offset (dungeon level 3 is for 5th level characters, and so forth).  Hmmm, that idea seems pretty good as well.

A related idea might be to skip directly to the level 3 action.  The characters would still have the chance to advance from level 1 to level 3, but all of that happens outside of the dungeon as the players get introduced to the setting.  They get some experience fighting bandits, help an old lady across the street, find and return Farmer Maggot's lost pig, the meat and drink of high adventure.  PC's wouldn't actually be ready to take on the dungeon until they're at least 3rd level, and then we follow a similar offset as above (dungeon level 1 is for level 3 characters, dungeon level 2 is for 4th level characters, and so on).

5E uses a lot of shortcuts from the older games - simplified stat blocks, for instance.  It seems like it could be amenable to a barebones treatment - big maps, sparse notes, lots of random content.  Maybe the 5E megadungeon can be as barebones as an old school dungeon; the referee just needs to be okay if a large part of the maps don't get used.  I'm hesitant about this approach because 5E seems to invite more set pieces and detailed encounter areas as opposed to improvisation - although it's not nearly as dependent on encounter design as 4E.  I supposed there's an implied constraint - we don't want to create too much content up front, but we want it to be deep or rich.

There really wasn't a 4E megadungeon to look back on as a reference, from what I can recall.  Because there was tight integration between the challenges and the player's capabilities, the focus in 4E was on smaller encounters and lairs that each culminated in a set-piece boss fight.  The nearest analogues to the megadungeon in the 4E world was Thunderspire Labyrinth or Chaos Scar.  Thunderspire Labyrinth was a mid-level adventure in a giant abandoned maze - it used a sprawling hex-crawl sized map.  In theory, the place was full of side passages and chambers that were unmapped, but the game product only covered a handful of lairs within the larger labyrinth.  The players typically got directions and guidance on how to find the lair as part of the plot hook or mission briefing, and it was assumed they stuck to the directions.  If you want to go big - like Mines of Moria big - then Thunderspire Labyrinth is a model that could work for 5E, too.  I shook the dust of 4E off my boots  a long time ago, but that was my favorite product from that period.

The 4E Chaos Scar was interesting because it created a game context for deeper = more dangerous, and I love those types of ideas as frameworks.  The Chaos Scar was a lair-filled valley, carved in the past by a falling meteorite.  Somewhere deep in the valley, the chaos meteor was still present, lurking.  The deeper the players penetrated into the valley, the more dangerous were the lairs nestled in the canyon walls.  It basically flipped the dungeon from a vertical construct to a horizontal construct.  Neat concept, but I didn't stay with 4E long enough to see how the final campaign played out.  The lairs were published in WOTC's  monthly online Dungeon magazine.  Does anyone know off hand if they were ever made free \ public?

I'm going to play around with encounter design and experience budgets over the weekend and see how the suggested numbers work.  I like the idea of collapsing the top levels down to a single level (covering levels 1-3) assuming there's a way to make it challenging but not instantly deadly - like having some gradients of encounters across the level.  I like the node-based lair approach of Thunderspire Labyrinth, too.