In my Pathfinder Tomb of Annihiliation Pathfinder campaign, my players experienced (for better or for worse) the Tomb of the Nine Gods, a dungeon designed by the lich Acererak. This tomb was written as a modern version of Acererak's other tomb, the famous Tomb of Horrors. In my campaign, both tombs exist, and my players have gone through them both. I was fascinated at how the players handled each dungeon, and I was able to see that both tombs served a distinct purpose in the game. While the new tomb exists largely as an homage to the original, and there are several references to the Tomb of Horrors, its design philosophy couldn't be more different. I don't think one can replace the other, and I don't think that was the intent, but the two dungeons perfectly demonstrate the evolving expectations of tabletop roleplay. Comparing the two reveals much about the flexibility of the tabletop RPG, but also how knowing your dungeon and selecting the "right" dungeon for your gaming group is important.
I run Tomb of Horrors about once a year, and have done for the past decade. It's usually done as a single 8-hour session on some special occasion, such as a 3-day weekend, to help players justify giving up an entire day to roleplay. It's a fun dungeon to run, and after you've run it 3 or 4 times you get to know it so well that it becomes easy to run. As a games master, a lot of it is puzzling and even frustrating. You'll encounter a sequence of "puzzles" with painfully precise requirements to "solve", only to discover that there's no penalty for failing and no reward for success. Many, if not most, puzzles are not clever or intuitive, and are instead opportunities for you to tell your players that their efforts have been in vain. It seems that the Tomb of Horrors isn't interested in creating a fun or horrific or terrifying experience. If I didn't know better, I'd probably guess that Tomb of Horrors had been written by someone who had heard of AD&D but had never read the rules or played the game. Of course we know this not to be the case, and I run Tomb of Horrors annually despite its design.
I'm not the only one to have been left unimpressed by Tomb of Horrors. Tomb of Horrors has a reputation among dnd players as "the most deadly" dungeon, and many people come to the dungeon with vague expectations of what they're in for. It could be a failing of me as a Game Master, but usually people's expectations are not met by the actual content of Tomb of Horrors. Almost invariably, players had imagined something worse than what they get, and something deadlier than what they just survived. If they didn't survive, then they imagined something cleverer or more brutal than whatever killed them. Tomb of Horrors has its own big shoes to fill but, through no fault of its own, is unlikely to ever live up to its own legend.
Tomb of Horrors isn't a very cohesive dungeon. There's only the barest story (a mean lich wants to kill you so you can't kill his mortal remains) to the dungeon.
I don't personally believe it was designed to be what it became famous as. That's not to say its designer didn't intend it as a deathtrap, but I'm not convinced its designer was trying to create a dungeon that would be played 50 years into the future through 5 major core system iterations. I think Gygax probably built the Tomb of Horrors for a very specific style of running and playing the game, because early rules simply didn't have the structure for repeatable experiences.
Put simply, Tomb of Horrors is not an exemplar of "clean" design or documentation, and while it's a great setting for a deathtrap, it does not itself succeed in creating a deathtrap.
When I think about a "deadly" dungeon, I'm actually thinking of a specific moment in some dungeon when my character became dead because of some ultimately severe event (probably, specifically, a trap.) My expectations of a deadly dungeon is that my character is placed into dark stone corridors, and then slain by a trap that was basically impossible to avoid. That's deadly. But it's also supposed to be a game, so before my character dies I want to have gotten past a few token traps to make me feel like my character wasn't entirely useless.
In my experience, that's been most people's expectations when they sit down to play Tomb of Horrors.
One of the best examples of Tomb of Horrors isn't the Tomb of Horrors module at all, but the Fighting Fantasy book Deathtrap Dungeon. This game book places you into a dungeon that's famous in-world for being a gauntlet of unexpected trials that slays, for the entertainment of the populace, everyone who enters. When you play the book, you expect to die and to have to go back to the beginning of the book and start all over again. Part of the experience is exactly that cycle. You enter the dungeon, you ignore the bait and leave a suspicious-looking treasure chest unopened, except you were supposed to open it so you take a wrong turn and end up dead. You enter the dungeon again, you open the box and get the reward, but you stubbornly persist too far down a different passage toward a volcano and you die of heat exhaustion. You enter the dungeon again and you open the box and take a right instead of a left and you're doing great until you end up in the hall of mirrors (and fellow players of this dungeon know what happens there.)
Each time you enter the dungeon, you last for as long as you can, and then you start over and try again using the knowledge you've accumulated from past attempts. I believe that's what most people want out of Tomb of Horrors, because that's how it's spoken of among dnd players. But Tomb of Horrors, as written, isn't nearly as deadly as most players expect from its reputation. Through trial and error, I'vo found that it is possible to make Tomb of Horrors a fun experience. As with any dungeon, whether "fun" means funny, deadly, terrifying, mysterious, or something else, is up to the group playing. Tomb of Horrors is ultimately whatever you make it, whether because the players try literally every stupid thing they possibly can, or whether the games master adds a bunch of new elements (threats, rewards, traps, and so on) into the dungeon to cause excitement, or whatever. As long as everyone's having fun, I think you're playing the dungeon as intended.
Acererak's latest dungeon, Tomb of the Nine Gods, isn't a deathtrap dungeon and doesn't claim to be. It is, in fact, a dungeon designed to be "solved." The puzzles in this dungeon are cleverly designed for reverse engineering. With enough experimentation, players are able to determine what a trap does, and then how to either avoid the negative effects or else distribute it among the party to minimise damage to any single character. Not every clue has a single definitive solution, so sometimes players are going to draw the wrong conclusion and take extra damage they didn't expect. With sensible pacing and careful study, however, players can generally navigate the dungeon without getting bored or dead.
It's a triumph of dungeon design. It maintains the anxiety of persistent threats, but also allows for exploration and interaction.
The question is, what do Acererak's 2 dungeons employ to create threatening (or even deadly) experiences that cause playful anxiety in players while also allowing for player victory?
A deadly dungeon must, given enough stupid choices, allow player characters to die. It's notoriously difficult (until it isn't) to die in D&D 5th edition, with Long Rests and Short Rests and all manner of healing spells and always-on Help actions, and Medicine checks and Death Saves. There must be a way to guarantee death in a dungeon that's meant to be deadly. (Personally, I'm a proponent of a cleverly concealed sphere of annihilation, because it leaves no remains to resurrect, but that's just my personal preference.)
The beautiful thing about sudden death is that it only has to happen once. When a dungeon demonstrates that instant death is possible, it's very suddenly a definitively deadly dungeon. It has been proven that it can kill. You don't need to belabour the point after that.
Some threats must not be "fair." A deathtrap that's obviously meant to be reverse engineered and avoided is not a trap but a puzzle, so sometimes the in-game designer must lie to the player characters. A clearly marked path around hidden pit traps must terminate with spikes that descend upon characters too busy poking the floor for traps to notice the one above. A puzzle that claims to hide a healing potion among poison must, in the end, turn out to contain only poison. A hallway collapses with age (it's not technically a trap!), a magical effect drains 10 hit points for no reason, and so on.
These "puzzles" are unfair but they're still solvable, if only by the law of averages. Some game groups will walk past a puzzle without taking notice, some will do the one thing that even the author didn't think of, and so on. Unfair threats are only unfair to those who fail, although admittedly they can be frustrating even to those who succeed because many players recognise the hand of fate when they see it, and that's discouraging for anyone trying to do the dungeon "correctly."
Then again, it also reaffirms that there is no incorrect way to survive a deathtrap. Sometimes, you also need luck.
Speaking of unfair traps, a deathtrap dungeon sometimes has an effect that gets no saving throw. There are established rules about what gets a saving throw and what doesn't, so I don't think it's defensible to just remove a saving throw because the dungeon is meant to be deadly. However, there are some effects that are so powerful, or so complete, or so sudden, that a saving throw doesn't make sense. Those need to be incorporated into a deadly dungeon, if only to account for the 1 in 99 player characters that manages to make it half way through the thing with 0 damage taken. There must be threats that even luck cannot avoid.
Monsters in a deadly dungeon must not be "balanced" for any specific level of play. Instead, monsters must be selected for story or theme. This is, in itself, one of the actually most fair ways of making a deathtrap dungeon deadly. Puzzles, traps, and magic can be fun and dangerous, but an angry hungry monster that happens to be 5 CR above what the player characters can handle makes perfect sense. In real life, if I were to go on safari in the wilderness, there would be no guarantee that I would absolutely never encounter a deadly snake just because I haven't qualified as a snake handler yet. Deadly beasts can happen anywhere, and they pose real problems to players. You can run, but it pursues. Maybe you can lock the thing in a room, but what if you need to get into that room later? Maybe you can use magic, but how long can you maintain concentration? How long does the spell last? Are there more of its kind that will come looking for it?
A random selection of monsters, of varying challenge ratings, is a great way to add threats that don't feel like deathtraps, but feel powerfully deadly.
I'm not one of those gamers who believes that survival is its own reward, or that it's fun to subvert expectations with empty coffers at the end of a dungeon. If players survive a deathtrap, then they ought to be rewarded. Imaginary gold is free, and a good deathtrap dungeon ought to have a lot of it, and probably a magic item or 2, by the exit.
The tombs of Acererak are excellent examples of drastically different design philosophies and techniques. Tomb of Horrors, in my opinion, is a setting you can use for your own deathtrap experience. Tomb of the Nine Gods is an exploration puzzle that's been meticulously written as a complete experience. They're both fun to run and fun to play.
Considered together, both tombs reveal much about dungeon design. Here are 5 principles for running a deathtrap dungeon, by my definition: