Wednesday, May 31, 2017

Megadungeons, Bosses, and Goals

In my last post, I mentioned that I dislike how Rappan Athuk used an "end boss" in the deep levels of the dungeon. The late levels (which I think were generated for publication) feature a delve that leads to the lair of Orcus. I think this is unfortunate, because it runs counter to not just the earlier levels of RA, but also generally to the philosophy that I think needs to guide the creation of a megadungeon.

Backing up several steps: megadungeons are best used for exploration-style play. This is why they work well with open table campaigns and, somewhat paradoxically, in convention play. Both scenarios were used early and often in D&D's development, much more often than the continuous party format that arose after D&D became popular among adolescent players with relatively stable peer groups.

With a continuous party of 3 to 6 player characters used consistently throughout the life of a campaign, going into small "module-sized" dungeons that take 1-4 sessions to clear, having boss monsters is fine. When they're in the Upper Lowlands Dungeon of Death™ they are doing it to fight the King Zombie, not because the ULDoD™ is interesting in itself.

Megadungeons are different. On a given delve, a megadungeon needs to be able to accommodate players who have spent 50 sessions going into the ruins, and players who are only going in this once. Maybe they're going together down to level 5A. If that's the case, level 5A needs to be interesting as an exploration goal in itself, without regard to whether the PCs ever go down to level 6A.

That doesn't mean either that every room in your dungeon needs to have a full array of what's interesting about it, or that dungeon levels shouldn't tie together in any way; neither of those is interesting. But what it does mean is that every level and sub-level needs to be a goal in itself, that it's worth going into it, and to be interesting if the PCs go there. All of this breaks down if the sub-level is just leading up to a boss monster. If level 8 is just a lead-up to the boss at level 9, the players who are only there for level 8 are cheated. And that becomes increasingly true as you get into the low levels of Rappan Athuk.

More than that, the boss monster is antithetical to the "living dungeon" concept of a megadungeon. By definition, once you beat Orcus or the Elder Elemental God or whatever, the dungeon is done. Subsequent expeditions are never going to have the same gravity as the one that killed Orcus. That kills the multi-campaign potential of the megadungeon dead. After all, you're putting this much effort into designing a huge dungeon, it should be good for more than one set of adventurers. (And having the next group kill Mecha-Orcus is worse because it just opens up an arms race of increasing absurd power levels that the OSR is pretty good at avoiding.)

There's more to get into with the living dungeon idea. At its core it means you restock and redraw maps, but it should always reflect the influence the PCs have had in some way. This is why there is a "vision and re-vision" component to megadungeon design. Done properly, the megadungeon becomes archaeological itself, with cues and remnants from past campaigns in future ones, and a richer experience overall.

None of this means that there can't be intermediate goals within the megadungeon. You can create a faction boss so that everybody remembers the time they fought and killed the Red Witch on level 6 – but that's one among many parts to the megadungeon's lore. You can have puzzles and ideas that span four or five levels at a time so the PCs unlock the Vault of Artasius on level 8C and find the Warhammer of Magnificent Smiting. But the campaign could go on after that, and there can be more intermediate goals. The megadungeon will never be fully cleared and there will still be mysteries for future groups to explore.

If you're committed to an exploration-oriented game, it should always be possible that the PCs never kill the Red Witch or open the Vault of Artasius. And it should still be a place worth exploring, and the players should still come away with memorable stories. It should even be possible for the players to find half the puzzles for the Vault of Artasius, and solve them, and then go over to a totally different path in the dungeon and never finish it. The megadungeon from this angle is really a commitment to sandbox-style exploration, with the dungeon as the "walls" of the sandbox.

This standard, where each part of the dungeon is interesting enough for a drop-in player but the parts work together in a way that is rewarding for the long haul, is the central design goal of the megadungeon. It's a difficult note to strike, and one that I don't think can be managed while designing with a final boss fight in mind. Which is why I'd encourage a megadungeon to not have an end goal, even though there are many smaller goals within its structure.


  1. "The late levels (which I think were generated for publication)"

    I think it is probable that, of the levels originally published in the module R3, only the following five were original:
    9: The Lower Temple of Orcus
    9B & C: The Well of Agamemnon
    10A: The Giant Cavern
    11: The Waterfall and Akbeth's Grave

    The other eight levels were, I think, created mostly for publication. All the levels in modules R1 and R2 were original.

  2. I am very much in agreement on most on these thoughts. But I don't think they are actually that specific to megadungeons. I think you can have a very similar style of game and campaign with multiple different dungeons on a smaller scale each.

    The approach outlined here sounds quite like "Many dungeons under one roof" to me. Giving each Dungeon Level it's own entrance and connecting them through "overland corridors" should lead to a pretty simimlar play experience, but I feel that such a campaign would lead to a much more believable fantasy world. With a megadungeon it's very obvious to everyone that the gameplay design was fixed first and then a fantastical world draped over it after the fact, with no attempt made to deny that such a place doesn't actually make any sense, regardless of how fantastical the world is made.

  3. What you says is true not just for mega-dungeons, but for regular campaigns as well. A Final Boss is fine in a video game where you stop playing after you've "beaten" it, but for an open-ended campaign it's better to have bosses stacked one on top of the other. If you want to turn one of these bosses into a Final Boss because it feels right at that time in the campaign you'll end up with a better story than if you pre-planned a boss as a final boss and the final fight was either too early or well after everyone was sick of the whole thing.

  4. That's a great point: each level must insist upon itself because there's no reason to believe that a particular player or particular character will visit adjacent levels. It is not merely a weird dungeon-as-funhouse nor is it the creator just showing off. It's necessary.

    You are very good at thinking through these things Wayne.

  5. Nice. I have to untangle these problems in an intellectual sense, but I've never had the thing arise sense since no campaign I've run has played out to a natural conclusion

    1. That's actually a good point in support of it. If you plan a conclusion for a campaign there are pretty low odds of the game ever getting there. Whatever kind of story you're adding to the campaign better has to work and be enjoyable to the players regardless of at which point they stop playing.

  6. Very interesting post Wayne. You've pinpointed for me why some dungeons sometimes labaled "megadungeon" just didn't seem "right".

  7. Very interesting thought-generating analysis, thanks!

    I think I understand the difference between goal/path campaigns and sandbox campaigns better now. The megadungeon you describe is a sandbox, whereas Rappan Athuk or Lost City of Barakus are ultimately goal-based campaigns that end when you beat the BBEG.

  8. I see where you're coming from, but I don't actually entirely agree. Without being able to speak to Rappan Athuk directly, I feel like every objection you have to megadungeon end-bosses on principle is one that can be solved through design and/or play.

    1. First, let's keep in mind why people use end-bosses at all: for closure. It's certainly possible to go into the Tomb of Mind-Blowing Flatulence to steal the Cloud Opal without fighting an elder air elemental... and as a rule people will go to the TMBF for reasons other than to fight the elemental... but for whatever reason (primed by simple narrative structures in action movies?) there are plenty of people who will find the dungeon more satisfying if the dispel the elemental before claiming their prize.

    And... even megadungeon campaigns come to an end. All campaigns come to an end. Maybe a group is breaking up, or the GM is moving away, or everybody's just tired of play-type A and wants to mix it up a bit. I see no problem with including a final level with a final boss for the people who want one.

    If, as you suggest, the boss warps the design of the dungeon as a whole (instead of just clues scattered around), that sounds like an execution issue. I'd say the ideal megadungeon boss is one whose influence can be seen through the rest of the place's structure in retrospect and who relates to the rest thematically, but who can be compartmentalized and only revealed when its time has come.

    You could argue that if it's just going to be cordoned off until a group-determined endgame, the GM might as well just slot in an end-boss on their own when they want it... and that's right for a homebrew campaign, but in terms of commercial products I'd say to err on the side of providing more material and saying "You don't have to use this if you don't want" rather than providing less and saying "If you want X, do it yourself."

  9. 2. I really don't see how an end-boss is antithetical to a "living dungeon." First, as above, its existence doesn't compel the party to stop exploration and try to rush the final confrontation in some sort of speed-run. They can and should explore to their hearts' content, aided by restocking, wandering encounters, and whatever else they'd be facing in the absence of a boss.

    And even if the party does go "Hey the enemy's gate is down" and charge straight through to the boss and "win," it's not like you can't use the dungeon. You don't even need to up the ante with a new, more powerful boss:

    * The focus can shift back to more leisurely exploration, or to extracting the dungeon's resources, or to colonizing it, or to dismantling the monsters' organizational structure. If you exile Morgoth to the void, maybe it's time to hunt down his lieutenant Sauron. If Sauron's power is broken, maybe it's time to see what the pretender Saruman is up to.

    * On that note, if you find yourself wanting a new boss, it seems best to restock the "boss slot" with something new. If the last boss was a combat juggernaut, replace it with something magical. If it was a magical beast, replace it with something insidious and mental or social. Make it a rival adventuring party. There are lots of things you can do besides just piling on HP and attack bonuses.

    3. The design principles you outline apply to end-bosses just as much to mini-bosses. I said this in a sense already, but just to be clear: the end-boss that you include in the dungeon is just like any other part of the dungeon: an option that's there in case the players want to engage with it, but not required. You can have an end-boss without tying them to a "save the world" plot. You can go further and make it clear that killing the boss won't solve the underlying issues or dispel the "mythic underworld." You can leave hints that the boss is load-bearing and that they should only face it when they're ready to be done, because after its death the dungeon will begin to decay and environmental hazards will increase in severity with each new delve.

    So yeah. Long story short, a megadungeon doesn't need an end-boss in a strict sense, but it can be nice to have one... and it should be easy to ensure that exploration-style "living dungeon" play is preserved regardless of when or whether the boss is ever faced down and defeated.

  10. You can have bosses *in* the dungeon but ime you cannot have a boss *of* the dungeon if you want to preserve it for indefinite future play. Inferior replacements don't work well, nor does escalation.


Comments on posts older than two days will not appear until approved.