Puzzle Difficulty?


Once again, I am a recent entrant to the IF community, which I apparently didn’t know about for most of my life. Yada-yada, you know the spiel.

Anyways, I was talking to someone about puzzle design, and how people come up with ideas. Most of the puzzles I make tend to have a similar style, which I call “engineering logic puzzles”.

The general idea is you have some system that accepts an input configuration, iterates on it in a way the player can observe, and then spits out an output. The goal of these puzzles is to figure out what input configuration gets you the correct output.

It’s really hard for me to come up with other puzzle formats, and these are usually fun for me to design, so it’s sort of a trope I maintain in all my projects.

However, the first adjective I got as a reply to this idea was “unethical”, to describe the predicted difficulty.

So, now I must ask: What would be regarded as the upper-average difficulty for puzzles in IF, and what would be regarded as incredibly-easy?

I understand if this is a difficult question to ask, but I’m a little concerned that my scale for puzzle design is currently set to “nightmare mode”, and I would rather meet player expectations whenever I can.



Well, this type of puzzle is a very broad category! If your system is, say, the SHA-256 hash function, then it’s going to be logistically impossible for anyone to figure out an input that produces a certain output. If it’s a logical XOR function (say, a light that’s attached to two switches), that’s pretty easy.

It all depends on how complicated the system is, and the logistics of making inputs and observing outputs.


There was a visual game I had made that demonstrates this idea, but it’s for Java 11 or something, and I haven’t made a web version of it yet lol. Otherwise I’d share it here as a demo for reference.

But yeah, you make a good point that this kind of puzzle mechanic definitely scales, and each scale level presents a very different difficulty.

This question was probably a bit too general, lol. My bad! :blush:


I think it’s like crossword puzzles. It’s all in cluing. Any puzzle can be very easy if the clues are really good, or really hard if they are underclued.

I think it depends on what YOU want. Do you want wicked hard puzzles for the hardcore puzzlers? Or gentler ones? You could have two modes player could pick from: hard and easy that would clue the puzzles accordingly.

But finding the sweet spot of puzzle difficulty is really hard. When I play games that do it well, I think “That’s an art form.”.


This is really good wisdom, actually. A lot of difficulty can be adjusted with the appropriate cluing and setting of expectation. I’ll be noting this down…!


Yeah, I think especially with these kinds of puzzles, it’s all about the intended audience (and framing expectations accordingly). Looking at graphic adventures, there’s a whole tradition derived from Myst and Riven where the fun involves futzing with complex devices and taking careful notes on how different functions change things; then there’s another tradition coming mostly from LucasArts and Sierra where the puzzles tend to the more straightforward USE X ON Y type, often involving some cartoon logic. Most modern IF draws more from the latter than the former, but there are definitely people who played and enjoyed the former games too (Myst is like, Andrew Plotkin’s jam, and So Far is very much drawing from it). So if you prime people that this is that sort of game, I’d think there’d be people who’d enjoy jumping into what’s a relatively rare kind of puzzle these days.

Of course, the flip side of that is that it might not have super broad appeal in one of the big comps. But what passes for “broad appeal” around here is like, a hundred people, and personally I think doing something idiosyncratic that’s going to be a dozen people’s favorite thing ever is often better (and more fun for the author!) than trying to chase an imagined audience. As always, testing is a key thing, but looking for testers who are going to enjoy this sort of thing rather than just throw up their hands at the very concept will probably be important.


I see you have found my mission statement! :smiley:


It was interesting for me to look at the past XYZZY Best Individual Puzzle winners and Best Puzzle winners. Not even playing the games necessarily, but also looking up reviews or walkthroughs and getting an idea of the puzzles:



Highlighting this not so much for your sake @inventor200 , but for everyone else reading. This can not be reiterated enough.

Although I have only tested a handful of games, I am proof that one person’s “easy” is another’s game-ender. In fact, in one case, I managed not only to miss a clue that all the other testers hadn’t, but to break the game trying to solve the final puzzle. More specifically, I tried a particular action at a particular point in the narrative that no one else had, uncovering a bug in the timing mechanism and rendering the game simultaneously unwinnable and unloseable. (It was as close to literally unending as a computer game can be.)


This is true. The question is: what’s the audience for the things you want to write? My first game I wrote for me. I was the audience for it (and look how that panned out!). Recent ones have focussed on beginners-- I really like the idea of writing games specifically to try and ease more people into the fold. My Spring Thing WIP is a pure puzzler-- I don’t know how big the audience is for that, but I always wanted to write one, so I don’t care if it doesn’t get a lot of play.

I think you can tell whether or not an author loved the game they made, and that makes a big difference in enjoyment. So write your super hard puzzles, let your testers guide you in cluing them, and see what happens. It’ll either succeed with the right people (yay!) or it won’t and you’ll learn something.


Start by making the puzzles make sense to you, then add some hints (which should be optional, naturally integrated in the environment or a mixture of both). After that, have some other people give you their opinions of the puzzles, preferably with you looking over at least one person’s shoulder as they play. Then tinker until you have a progression that makes sense.

In general, if a puzzle is deemed too difficult, there’s a few things you can do without changing the core of the puzzle itself:

  1. Put it later in the game.
  2. Add more clues.
  3. Include an antepiece - another puzzle that has to be completed or at least meaningfully engaged with before this one must be solved, that introduces the idea(s) that are causing confusion. If several aspects of a puzzle are causing problems, these can be separated into multiple antepieces or combined into a bonsai (less complicated) version of the troublesome puzzle.
  4. Move it somewhere geographically where the player is more likely to be thinking in a manner that will help solve the puzzle.
  5. Add an alternative method of reaching the goal (which may mean making the puzzle optional-but-helpful, or going systematically through the object list to see what else would sensibly help, or reconfiguring the puzzle a little to add the alternative - the latter, admittedly, does change the puzzle but preserves the original solution, which is sometimes desirable).

It’s common for games to have one or two difficult puzzles, but I don’t think all the puzzles should be difficult. It’s preferable to have puzzles with a range of difficulty from easy to hard. As a general rule, the easier puzzles should be earlier in the game so that players feel like they’re making progress and they don’t get fed up and give the game away before they’ve really started. In other words, suck them in.

It’s also nice for puzzles to be multi-layered, so you have to do A before you can do B before you can do C. A series of easy puzzles may lead to the solution of a hard puzzle. The hard puzzle only appeared to have been hard because you hadn’t yet discovered all the pieces you needed to solve it.


I LOVE it when games do this!! Noted!


I think this is great design advice, I’m going to remember this for the future.


I agree! The one wrinkle I’ve also heard folks propose is to make your hardest puzzles the penultimate ones, so that the very final challenges are a easier and the player has strong momentum carrying them through the climax of the story. I think this can go both ways, though, since if the game is more puzzle-focused it could feel disappointing if the final set are relatively ho-hum. Possibly one could split the difference with an elaborate-seeming, highly complex puzzle that’s nonetheless not too difficult to solve? I’m thinking of the Boiler Room in Cragne Manor as an example…



Oh, so kinda like a victory-lap sorta task?


Yeah, that is exactly the phrase my brain wasn’t coughing up as I was writing this!


I also think that the last puzzle works best as post-climactic descending action, though there have been some very exciting and memorable exceptions (Spellbreaker)!