DMing Puzzles: Not Too Easy, Not Too Hard, What's Just Right?

[OnlineDM]

I'll also share that my initial version of the Room of Runes puzzle from Descent Into Darkness was too hard. It was one of those things that totally made sense if you came upon the right way of thinking, and was otherwise just not going to happen. The current version is sadly somewhat similar, but I've provided the players with hints that they can earn.

A great puzzle is one that the party can actually work on and get closer to the answer. A not-so-great puzzle is one that is completely opaque until you guess what the GM was thinking, at which point it's solved with no additional work. The latter type is really a riddle rather than a puzzle. Riddles can be fine in RPGs, too, of course, but if you're going for a puzzle and you create a riddle (as I think happened in the OP's case), then you're probably not going to be satisfied with the result.

 

[Stormonu]

Judging whether a puzzle is too hard or easy is a difficult thing; it's all relative to what the Players (not their characters) have run into in the past. I've run across puzzles that have just clicked for me, others that I'd never have figured out without help (Yours above would have been one that I don't think I would have solved without clues).

I agree that when designing a puzzle, you should look to include around 3 clues to help make it easier for the players to solve. How they get these clues (found in a book, Int checks, skill checks) is up to the DM, but they should there to help the game move on if the characters just don't grok the puzzle (and it's important for them to solve it to move on - if it's just a side puzzle, clues are optional...). Though I am loathe to do it, I've pitted parties against puzzles I had to somewhat metagame giving clues out for them solve. It happens, especially if the original "riddle" turns out to be too vague (or you do like I once did and screw up the math in the puzzle...).

Personally, I have no problems with lowering XP from the puzzle if the players don't use their own brains to solve it or reducing XP for each clue given out (advancement can be its own reward).

 

[Jhaelen]

And when I say 'realistic,' I mean a puzzle that any mage worth her salt might invent to protect her magical laboratory and doodads. All it takes to make a tough puzzle is two or three layers of misdirection, like I did with this one.

Actually, I believe the most realistic puzzle is to not use a puzzle at all. If I was a super-genius mage, I'd make damn sure to use a mechanism that no one will ever be able to figure out or circumvent. Putting clues somewhere, no matter how cryptic, is _stupid_ if I don't want anyone to figure it out.

Likewise I never understood why magic items aren't all custom-made to only work for the intended recipient (and naturally its creator).

So why would there ever be puzzles or magic items that anyone could use?

Because it wouldn't be fun otherwise! And that's why you shouldn't try to come up with a 'realistically tricky' puzzle but with a puzzle that your group of players might consider fun.

Personally, I'm very reluctant to put puzzles in my adventures because usually even puzzles I consider to be very easy completely flummox them. They simply aren't that much into riddles.

I have the opposite problem when I'm the player, though: I tend to breeze through any kinds of puzzles my fellow DMs come up with, because I _am_ into that kind of thing. It's just not fun for anyone else.

 

[delericho]

Actually, I believe the most realistic puzzle is to not use a puzzle at all. If I was a super-genius mage, I'd make damn sure to use a mechanism that no one will ever be able to figure out or circumvent. Putting clues somewhere, no matter how cryptic, is _stupid_ if I don't want anyone to figure it out.

Exactly. Unless the mage in question is particularly forgetful, the sequence is just going to be a random combination which he'll then remember.

Actually, I recall one of the Exalted pre-gen adventures was built on this principle - the dungeon was filled with traps, secret doors, and puzzles... but all the clues were wrong because the in-game designers had recognised the silliness of giving out clues for people.

(That said, in my experience most IT guys seem to have a couple of "password" stories - stories of otherwise-intelligent people who put their password on a post-it note on their PC, or set the password to "password", or both. So perhaps it's not all that unrealistic...)

Likewise I never understood why magic items aren't all custom-made to only work for the intended recipient (and naturally its creator).

So why would there ever be puzzles or magic items that anyone could use?

Because it wouldn't be fun otherwise!

That's exactly it. Puzzles aren't realistic, but they are (or at least can be) fun - they're there to challenge the players, not the characters. (And, besides, the 'realism' argument is a bit weak when you stick it up against everything else in the game.)

Anyway, regarding the actual difficulty of puzzles, I recommend a model I observed and then stole from the Lego Star Wars games:

In Lego Star Wars, each level has three rings of puzzles.

In order to complete the level, there tend to be one or two small and easy puzzles to solve. A door needs opened, so you have to figure out how... but that's not hard. Or you need to use the Force to move the plug that's sitting right there into a socket that is also sitting right there. Easy, simple, and dealt with quickly.

However, there are then various optional sub-quests you can complete within the level, if you are so inclined. Basically, the levels each have ten 'canisters' hidden throughout - you don't need these to proceed, but if you are so inclined you can spend more time digging these out. (In LSW, you tend not to be able to collect all of these in a standard walk-through - you have to revisit the level later in 'free play mode'.)

And then, finally, there are a number of Easter Eggs hidden in the game, so that people who are really obsessed can spend ages just exploring, visit every little nook, and there are things that are placed there just for them, so they can say "cool!".

Applying that to D&D, then, I would suggest the following:

To complete the quest(s) that were the party's motivation for going to the dungeon in the first place, there should be a small number of fairly easy puzzles. These may well not be true puzzles as such at all - if there are multiple routes through the dungeon that go to different places, merely finding the right one may be the puzzle. Either way, keep these puzzles few in number and easily solved.

In addition, the dungeon could (and probably should) be laced with a number of other puzzles that are not required to solve the dungeon, but give access to additional treasure, or allow them to swing an NPC to their side, or whatever. And so, players who are interested can spend a bit more time dealing with these puzzles, but players who aren't interested or can't solve them can just move on without losing out too badly. Being optional, these puzzles can be significantly harder; there should probably be a range, though, from "quite challenging" to "fiendish".

And then, finally, there could perhaps be a few other puzzles that serve no other purpose than to just have players say "cool!" Like when they realise that the dragon that was referenced in their very first adventure is actually the same dragon whose bones are used in the giant's throne this time. Or something.

(Incidentally, to the OP: In general I would caution against tying a trap to a puzzle. It's frustrating enough to get the wrong answer; it is considerably worse to get the wrong answer and get zapped for it. At worst, I would suggest a "three strikes, then lockdown" approach to the puzzle.)

 

[Nagol]

Actually, I believe the most realistic puzzle is to not use a puzzle at all. If I was a super-genius mage, I'd make damn sure to use a mechanism that no one will ever be able to figure out or circumvent. Putting clues somewhere, no matter how cryptic, is _stupid_ if I don't want anyone to figure it out.

<snip>

The time I use puzzles is not when the creator is expecting to be the only one to enter an area: it's when the area is meant for a larger semi-anonymous clique like a secret society or cult.

The puzzle acts as a relatively dumb gatekeeper -- keeping out those who cannot prove they are meant to be allowed in. Th answer to the puzzle is something that a member of the group is expected to know or be able to unravel and that others would struggle with.

 

[delericho]

The time I use puzzles is not when the creator is expecting to be the only one to enter an area: it's when the area is meant for a larger semi-anonymous clique like a secret society or cult.

The puzzle acts as a relatively dumb gatekeeper -- keeping out those who cannot prove they are meant to be allowed in. Th answer to the puzzle is something that a member of the group is expected to know or be able to unravel and that others would struggle with.

That makes a lot of sense.

Unfortunately, such a puzzle would not, realistically, be something one could puzzle out, but should rather be tied to some specific detail of the sect. ("The word of God - only in the footsteps of God will he proceed." - easy if you know the 'name of God', but not something you can work out without it.)

The problem there is that it probably isn't reasonable to expect players to know the ins and outs of semi-secret societies in the game - even if you give them the appropriate handouts and/or unfettered access to the books, who really wants to do homework for a game? And so, it comes back to character knowledge, which tends to boil down to skill checks.

Not that I have anything against skill checks... but they tend to be rather dull, and a poor way to handle puzzles.

 

[jedavis]

We ran into an awesome puzzle in a game I played in last summer; the DM knew that we knew how to puzzle out this particular thing (cellular automata), and then built off of that. Now I want to try Turing machines, finite automata, graph colorings, and max-flow problems... The joy of gaming with programmers.

I also agree that hard puzzles should be used sparingly and in the "bonus!" mode, where solving the puzzle generates extra treasure, but failing does not prevent achievement of the primary objectives.

Another thing I've looked at doing is stealing puzzles from Microsoft's Puzzle Challenge archive... but my players played in it in previous years, so I'd probably have to go pretty far back. A lot of these are a bit too systemsy (gonna have a hard time justifying assembly language in-world), but some of the more theory and wordplay oriented ones might work.

 

[Nagol]

That makes a lot of sense.

Unfortunately, such a puzzle would not, realistically, be something one could puzzle out, but should rather be tied to some specific detail of the sect. ("The word of God - only in the footsteps of God will he proceed." - easy if you know the 'name of God', but not something you can work out without it.)

The problem there is that it probably isn't reasonable to expect players to know the ins and outs of semi-secret societies in the game - even if you give them the appropriate handouts and/or unfettered access to the books, who really wants to do homework for a game? And so, it comes back to character knowledge, which tends to boil down to skill checks.

Not that I have anything against skill checks... but they tend to be rather dull, and a poor way to handle puzzles.

Depends on what you choose. For example, for the same source material "Only a penitent man..." to get down on your knees doesn't involve a lot of esoteric knowledge. It's just a veiled instruction couched in the society's language much like "Speak friend and enter".

A lot of puzzles can be tied into a more esoteric field of knowledge, but still have an internal consistency that an observant or committed group can work through the logic. Puzzles of this type I've used include magic squares for a numerologist college, nonsense written in one alphabet that makes sense when spoken aloud to a listener of another for a bardic campus, and the selection of Fibonacci numbers from a sequence for a secret society.

Of course, without an observant or committed group, the skill check still exists as a safety valve.

 

[kitsune9]

My take is that if I'm going to invent a puzzle for my players to solve, then I'm going to create one at their ability to solve. My players are not puzzle-solvers. They don't like riddles and puzzles so when I throw one at them, then it's got to be a simple one or nothing at all. Granted it's not realistic of the high-level 26 Int evil lich or wizard to create a simple logic puzzle, but if that's what I got to challenge my players, that's what I'm going to use. No sense in them wasting a couple hours and getting pissed off in the end anyways.

My litmus test is to design the puzzle and give it to my non-gaming wife to solve it. I time her so I'll know how long it will take her. If she goes longer than five - ten minutes or solves the puzzle incorrectly, then I go back to the drawing board.

Back in my 2e days, I had two engineers and a nurse who loved puzzles. I had to pick out puzzles that I would have no shot at solving (let alone approaching correctly) in order to challenge them. They still solved them in less than 30 min.

 

[Jhaelen]

The time I use puzzles is not when the creator is expecting to be the only one to enter an area: it's when the area is meant for a larger semi-anonymous clique like a secret society or cult.

I agree that there may be situations that require you to leave clues to allow fellow members of your secret organization to solve it. There are real-world examples for this.

One of the most common involves a common reference known only to members; this could e.g. be a book, a song, or a prayer. So, to be able to make sense of the clues, you need to have access to this common reference. If you'd like your players to solve the 'riddle', they'd first have to figure out what that common reference is and get hold of it.

I'd like to use this opportunity to point out one of my favorite novels that deals with this kind of thing: Cryptonomicon by Neals Stephenson. It features a very interesting cipher that can be decoded using a set of playing cards. It's a good example for a 'riddle' that is impossible to solve for anyone who doesn't know the method to decode it but easy (if time-consuming) for anyone in the know.

Of course once you start to introduce magic in the equation, all bets are off. You can use magic triggers to make sure no one except members of your organization will actually receive any clues by adding requirements that are difficult to match without being a member. Even something as simple as an alignment-based trigger would make it hard for non-members to get to the clues.
Everything detectable by magic can become a part of the trigger.