How do I design a puzzle where players are supposed to fail 4 to 5 times before they succeed?

Question

I want to introduce an omen of an incoming demon incursion to my party that threatens the world my PCs live in. My plan is to have the party gradually trace down clues / investigate deaths and murders before finding out the truth and (hopefully) stopping it.

One "clue" I want to give to players is a magical item that materializes only when evil is approaching, However I felt just giving them the item isn't impactful or foretelling of anything. I can't verify if the players would wonder "Why did we get this weapon?" and link it to the other clues I'm handing them. Unless I wrote on it "EVIL IS COMING!!!"

My solution to this is a magical item that "follows" the party. Revealing itself as a recognizable puzzle multiple times. One that's easy to fail for the first few times but on the 4th or 5th attempt the item is given to the players. I intend on the party thinking that the item found them.

I plan the "puzzle" to appear in a variety of places across the campaign: in a sewer, a dungeon, an empty room of a manor to name a few. To tell the players that this isn't some strange dungeon specific thing, but something with more intrigue.

This, I hope at least, combined with many of the other clues (I'll hint to players of this magical item and its properties in other places) will give players enough information of the incoming incursion. But I don't want players to give up part way through and skip the item entirely because the puzzle was too difficult...

So my question is:

How do I design a puzzle where my players are intrigued enough to solve the puzzle and aren't too discouraged by failure to receive the magical item?

Footnote: 4 to 5 times is a general approximation. I feel any more or any less would make the puzzle too easy / too hard but if anyone want's to argue more or less by my guest.

Answer 1

There are a couple of common, tested solutions to this kind of problem, but first, since you discuss a "puzzle", I'll note that I've found it tricky to implement puzzles in tabletop games. I'd recommend against blocking success in the campaign behind a puzzle. Here are a few techniques I've used, including one that still includes puzzles but mitigates their potential problems. The big takeaway that they all share is that you should design your story so that it doesn't grind to a halt if the players are unexpectedly bad (or good!) at solving your challenges.

Multipart Plot Coupons

TV Tropes calls this "Gotta Catch Them All": there are a set number of items (or spells, or people) that the players have to assemble in order to gain access to the magic item. Instead of needing them to fail the first several attempts, they could find a clearly important or mysterious item in each location that you encourage them to take with them (maybe by making them seem valuable or tantalizing in some way).

In the most basic, video-gamey sense these can be four elemental crystals or the pieces of a shattered object, but I've run campaigns where the players had to find and protect the keys to magic seals that needed to stay closed, or were following the steps to a prophecy that they didn't fully understand.

Some Assembly Required

Another approach that comes to mind is to make a magic item that comes in functional parts, or evolves as you use it. The classic D&D example of this is the Rod of Seven Parts. Each piece is a useful magic item on its own and you might not realize that it is the part of a whole or how many parts there are in total. However, when you find an additional part and connect it, the artifact becomes more powerful.

An example of an evolving artifact I've used is Faarlung's Algorithm from 4e, which is a magic orb that's also a weird puzzle like an arcane Rubik's Cube. You can spend time trying to solve it (making hard Intelligence checks) in order to make it more effective. Perhaps the players can find clues or materials in their journey to enhance the item.

The key to this approach is to make the item immediately useful, even if it's not particularly strong. That way, they'll keep it with them and you can gradually bring it up to its full power.

Rivals and Conspiracies

Lastly, you can make the actual solution of the puzzle less essential to progression. Make it so that, if they solve the puzzle before you want them to get the true reward, they get a minor reward but the goalpost gets moved. Conversely, if they fail to solve the puzzle, someone else collects the reward, the players recognize it's something valuable, and they get a lead to chase it down.

This can be as simple as a rival group of adventurers or villains that's after the item for different reasons. Think of the Indiana Jones movies: the protagonist has rivals trying to get the same artifact, and there is a push and pull where they each get part of the final reward and chase each other around to get the whole thing. Your rivals could steal the item just as the puzzle is solved or, conversely, brag about solving it themselves after the players fail.

You can do this without an antagonist, too, by including a friendly character who keeps secrets. I've had great success with the classic "secret dragon" gambit, where a powerful and mysterious character has the PCs go on missions to collect things. The players slowly get suspicious until it's revealed that the character is a dragon (or some other powerful figure) in disguise. Each puzzle attempt could end in the players turning over the item or a clue to it to the benefactor, or if they failed, the benefactor could come in afterward to collect it. When the players become suspicious enough, they can confront the benefactor, and it could culminate in a fight or a friendly confession, as appropriate.

Answer 2

You can use a combination puzzle - but don't decide the successful sequence in advance.

I have used a puzzle dungeon, where the party were absorbed into a mysterious cube. The dungeon was cube shaped but appeared flat to the players (initially). Soon enough the realised they weren't going in circles but were on a cube demiplane. They had to find a certain combination in order to escape the cube. There were some clues in the way of colours and shapes.

I let them investigate and begin to find the "right" combination. Of course, they did not know that I was changing the combination. So, instead of having a magical item materializing, I put a room on the dungeon where there was a door which they could not reach. There was an impenetrable energy dome preventing passage. They figured out that the energy pylons were feeding this. That's where the fun began. There was a system of colours, shapes and switches dotted around the whole dungeon which they had to try out to see if it turned off one of the pylons. I put 3 in my puzzle.

I did not decide the exact combination in advance. Instead I let them theorise and explore the dungeon, to try different things out. I went with a couple of their theories: one was to do with egg-shaped objects. The tried it out and I thought: "Why not?". So, as they explored they found these, including one being an egg-shapped eye from a beholder. Obviously, I added a few traps, set encounters and random encounters along the way. Also, as I designed it as a cube dungeon, I had a theme for each side, e.g. one side was about (un)death. I used davesmapper to generate the cube. The rest was deciding what was in each room.

It took them around 3 sessions of 4 hours to escape the dungeon and they had lots of fun.

As, I mentioned, it can help if you set the puzzle as a task for finding a given combination. You decide the combination as they go along. With the magical device, maybe once they find a correct sequence, it could light up. That way they know they got one part right. This way, you decide how many sequences they need to get. In my example, if I did that, I might have removed the 3 pylons and instead suggested that the impenetrable barrier was getting weaker gradually with each successful "discovery". You decide when enough is enough and they find the winning combo - which you basically decide whenever you want.

Answer 3

A plot-locked puzzle is a bad idea, and it won't achieve what you want it to.

In books or movies, or video games that have a forced linear story, the author could do what you are suggesting. The protagonists receive clues over time, but they don't put them together until the author decides it is time for the mystery to be revealed. Or a protagonist continuously works at a puzzle, but they don't solve it until the plot demands they solve it, at which point the puzzle points them in the right direction.

This doesn't work in a tabletop RPG because it's an interactive medium. Instead of puppet protagonists who follow an on-rails script, you're dealing with human players who want agency and who expect to make meaningful choices. An unsolvable puzzle is antithetical to this, and would probably just frustrate the players and make them dissatisfied with the campaign.

The notion of a puzzle that isn't solvable until X number of guesses, this doesn't make a lot of sense. What if the players guess correctly the first or second time? If they guess wrong once, why would they want to keep guessing? A puzzle's difficulty should not be measured by the number of expected incorrect guesses before a correct solution is viable. That's not how puzzles work.

From your question, it seems like your actual goal is to drop clues at dramatic story moments, and have these clues eventually warn about an incoming invasion (or some other complication that the GM knows but the players don't). This is a fairly common trope, and there are better ways to handle this than a bad puzzle.

(1) If you want to stay in control of the campaign story, and keep things mostly on-rails, then the clues could be presented as pieces of a larger whole. Gregory Avery-Weir's answer describes this well; maybe each clue is one piece of a set, or a message fragment that can't be read until all the fragments are gathered. Clues are presented in a way that communicates they are not the complete picture, and that the players can estimate the total number of parts. This gives the players a clear sense of progress as more clues are revealed, but also informs them that they probably can't short-circuit the GM's intended story progression.

(2) Alternatively, if you trust your players to steer the campaign direction, then present the clues, and let the players interpret the clues however they wish. Maybe the players will correctly interpret the clues earlier than you expected, and you as the GM would need to figure out what happens next. If you want the players to deduce the clues before the campaign can continue to the demon incursion arc, you should present at least three clues to mitigate factors like missed clues, failed checks, or misinterpretations. This approach may be more fun for the players because it lets them flex their player agency, but it also puts more pressure on the GM to improvise.

Answer 4

Write on it “EVIL IS COMING”

I want you to think about popular media where there was a BBEG that posed an existential threat to the world. Think Game of Thrones, the Lord of the Rings, the MCU, any James Bond film, Sean of the Dead, I am Legend, the Day after Tomorrow, you get the idea.

I want you to think carefully about when the audience, not the protagonists knew the stakes. Pretty early on, right?

While it’s fun for us as the audience to know that we know what the protagonists don’t know, it’s no fun for the protagonists. The fact that it’s no fun for the protagonists is part of what makes it fun for the audience.

In a TTRPG, the players are both audience and protagonists. Trying to play a TTRPG when you don’t know WTF is going on is not fun. You do want your players to have fun, right? Don’t treat them like they are just protagonists because they aren’t.

In any event, most players can’t read anyway.

Answer 5

My Experience

I never went through the exact situation you are asking about, but I think my experience transfers correctly here.

I was one of the 4 players, in a group where we all knew each other. Our characters had some kind of "rollback in time joker". We had been trapped in a room with a strange device.

After some time passed, the device would move and some kind of laser would hit a target, then the room would be engulfed in flames, killing us and triggering the time rollback.

The solution was to move a specific set of lenses in the path of the laser so that when it hit the target it would open a path that we could take to flee. The combination wasn't random but it was very hard to guess so we did a fair amount of trial-and-error before we understood the solution.

The main difference with what you are attempting is that that was a life-or-death situation and it could have ended the campaign if we never found the solution. In your case you don't have that issue.

Make them fail

You want your players to fail (at least the first few times), so there has to be a clear failure condition. I recommend using a hidden timer (something like "ten minutes after someone has entered the room, the puzzle teleports away"), that way you are almost guaranteed that the players won't figure it during the first encounter.

The second time they encounter the puzzle, they will have their previous experience to help them and know they have to be quick (or will suspect that something they did wrong the first triggered the teleportation time and they will be extra careful). That said, they probably didn't think about this puzzle in the meantime so unless your puzzle is easy (it shouldn't) they will fail once again. From then you can expect your players to guess this puzzle probably will come up again later and try to figure it out before that happens.

Let them succeed

Failing may be fun, but you want them to eventually overcome the difficulty! Depending on how good they are with puzzles in general it may happen "naturally", or they may loose interest by the third encounter. However you have a big advantage here compared to most puzzle situations: your players have time. They may think about this puzzle between sessions, or during the lunch break, or any time the DM is taking a bathroom break...

On top of that, there are ways for the PCs to find more information in-universe: maybe they know someone who is fond of puzzles who can try to help them? Even if they don't think about anything you can give them clues in an old dusty book found in a dungeon, or by having an NPC talk to them about the topic the puzzle is about... All those tricks feel cheap when used in solve-or-die kinds of puzzles but here you can take all the time you need to make it subtle.

One Example

What follows is only a example, that follows the previous points.

The Lost Crusader was a very powerful Paladin, who killed hundreds of fiends with his legendary blade Hornslayer. Unfortunately some devils tricked him into signing a pact and he died dishonored. For his soul to recover peace, his sword has to be wielded one last time and be used to kill those specific devils, without failing to respect the pact.

The ghost of the Paladin is trying to give his sword to the PCs, but the pact makes it so that he can only appear in the ten minutes that separates day from night at dusk. He also can't speak. For someone to recover the sword they have to kneel and recite the oath of the fiend slayers

During the first encounter, as they are setting a camp the PCs will see a silent ghost who appears to try to hold his sword as if he was trying to give it to them. If they try to take it their hand goes through, and the ghost doesn't seem to understand them. As the last ray of the sun disappears, the ghost vanishes.

A week later, the same ghost appears, and does the exact same thing. The PCs may try to speak to him in different languages, to trade him something, to inspect him... Still no luck so far.

The next day, as they pass by a town, they ask a few of the townsfolk about the ghost, and among various nonsensical rumors someone suggests the ghost may be of a member of the fiend slayers. As the PCs visit the Temple of Fiend Slaying they can see that all the fiend slayers have this same insignia and they don't recall seeing it on the ghost. That's because the ghost is old, the order was founded by him, and only after that they started to wear the crest.

Ten days later, the ghost appears again, and they look for the insignia. They don't find it but notice that the sword looks like the sword on the crest and the helmet has the same shape: clearly there is something fishy here. Still they don't have more idea.

Later on, they rescue Malia a young gal and she tells them she is an apprentice of the Order of Fiend Slaying. She asks them to bring her to one of their temples. As they remember the ghost they ask people there and learn about the Lost Crusader: there are too many similarities for this to be a coincidence! Still they don't know what they are supposed to do with the ghost. A monk suggests that maybe they should try sparring with the ghost (that's a red herring).

Even later, the PCs receive an invitation for Malia's graduation: she is going to swear her vows and become a true Paladin. They come and see the ceremony. At some point her master offers her a sword, she kneels, swears, and receive it. The way the master holds the swords looks a lot like the posture of the ghost.

The next time the PCs encounter the ghost one of them -who was already ready to spend the rest of his life slaying fiends-, decides to kneel and swear. The ghost then gives the sword and vanishes (he actually still inhabits the sword for as long as the devils are alive).

Of course, the PCs could not have got the last clue and still be stuck, but as the DM you always can give more clues, up until the point it becomes very obvious. You should also offer them other ways to solve the situation: let's say none of them are fond of swearing as they don't think they will be able to hold their word, maybe they can convince a paladin NPC to travel with them and swears.

Answer 6

Players will find hints where you don't put them, Direct them to the mechanic at play

If the sole mechanic preventing the party from obtaining the item is that they "must try 5 times" then illustrate that requirement.

Have the item appear, a clock of sorts begins and starts counting down. They can poke, prod, what ever but investigating the item will be the most important action to take.

The clock will count down regardless, and in that time they may notice 3-5 empty slots the first time they see it with one slot either physically closing or filling with light, a magical key is forming, some sort of state change that occurs over the duration. (A measure of determination to get it open if it must make sense lore wise)

On the 2nd visit, that slot will be lit up or filled or closed, then the 2nd one will begin doing the same.

Once the last failure occurs, the next visit, all slots, lights, keys are visible, to which a new occurance can happen. All the keys turn (a reward for never giving up and staying determined to solve the riddle) or what have you to reveal the item then actually need and can get a hand on.

It may be a prevention for passerbys being targeted and getting the item without staying in the area long enough, or a frustrated BBEG who might given up after the 3rd try "It's impossible and it's locking me out even more as I try!" etc. preventing them from ever getting past the 3rd key, reverting to 2 on their next attempt.

Possible requirement: They must still be trying to solve the scenario while the clock runs out each time to count.

Answer 7

While the other answers are probably right on how you should do it (ie. not be an actual puzzle, have separate parts etc.) having just one puzzle is still possible.

Wordle, for example, almost always takes 3-5 attempts to solve. If each time the party sees the puzzle they are able to make one attempt at it, and each attempt gives new insight for the next attempt, you can make it very unlikely for the puzzle to be solved the first time.