Best Puzzle Practices

Hi, this is regarding a text adventure that uses a parser, so while it does have visuals the interactions are all in text.

I’m somewhat new to the community in that I started writing an IF game using Inform back in 2005 or so, never finished it, dusted it off in 2015 where I actually posted a couple questions here, had fun reminiscing with it then didn’t start work on it again until 2025.

All this to say, as I went through the original game I realized two things; the map was far too empty, and the puzzles for the most part followed the patterns:

Use X to get Y
Use X to bypass Y and open access to Z

There were some exceptions, but as I update everything, I’ve been making a point to upgrade the puzzles and make them more interesting. I’ve gotten a lot done but am still working out some sections so I’d love to hear people’s takes on what are some puzzle types you enjoy, types you hate, other variations beyond those simple two I mentioned, and also when adding more complex puzzles, how complex is too complex?

2 Likes

Welcome!

IF authors seem to like making taxonomies of puzzles almost as much as they like making puzzles. A few I’ve found useful:

Graham Nelson - Inform 6 DM4:
Mazes • Light source puzzles • Capacity and exhaustion puzzles • Timed puzzles • Utility objects • Keys and doors • Machinery and vehicles • Fire • Water • Air • Earth • Plants • Animals • Monsters • People • Ropes and chains • Riddles • Decipherment

Dennis Jerz - Puzzles in Interactive Fiction:
Unlocking doors • Manipulating objects • Recognizing patterns and loops • Navigating mazes • Interacting with characters • Guessing the verb

Ryan Veeder - On Puzzles in Interactive Fiction:

  • Does the player need to know something?
  • Does the player need to try something?
  • Does the player need to notice something?
  • Does the player need to be creative?
  • Does the player need to apply logic?
  • Does the player need to do something that seems self-destructive or undesirable?
  • Does the player need to do something else first?
12 Likes

This is great, exactly what I was looking for, thank you!

I’m fond of EJoyce’s metric of hammers and light bulbs (see that thread for more; they borrowed this metric from a puzzle hunt they did, and it’s in response to me proposing a couple types of puzzles…everything is connected).

They suggested that you can think of puzzle difficulty both in terms of “light bulbs” (how much of a leap of logic does it take to discover the solution?) and “hammers” (how much straightforward effort does it take to implement the solution?). Different people have different tastes on this, so many games have a mix of both: exploring and mapping until you find the thing you need would be a hammer-heavy puzzle, while realizing that thing can be used to bypass an obstacle could require more light bulbs.

My philosophy is that adding more hammers is easy but adding more light bulbs is difficult, so my advice will focus on the light bulbs. (You can always make your map bigger, for example, if you need more hammers, or break your key item into two pieces and hide them in different places.)

I think it was Dan Fabulich who said that the best light bulb puzzles, like the best jokes, are obvious in hindsight. Once you’ve realized the answer, it’s clear that it has to be the answer, even before you type it into the game. For that to happen, the player needs to have a clear understanding of what they’re capable of in the game—what actions are possible and what those actions will generally do.

My favorite way to get this clarity is systematicity. The player is able to do some unusual thing, and that thing follows predictable rules that you can try everywhere and learn via trial and error. In Scroll Thief, there’s a summoning spell that will teleport any living person to your location, but it requires line of sight. How can you use this to your advantage? Well, perhaps there’s an object that can’t be carried through a certain door, or an alarm will go off. Have your friend pick up the object, then look at them through the door, and summon them to you. Everything here follows clear and learnable rules that work throughout the entire game (you can summon a person with line of sight, you can see through open doors, you can tell someone to pick up an object and they’ll do it); it’s the combination of those regular rules that creates the light bulb puzzle.

Sometimes you don’t need this kind of full systematicity. Sometimes you can rely on players’ existing knowledge of the real world. If there’s a piece of cheese and a mouse that needs to be lured out of hiding, it doesn’t matter that nothing else in the game gets lured in by cheese—that’s a cliché about mice in the real world that many people will know. That can also make a good light bulb puzzle; just make sure the knowledge needed is really universal and not specific to your culture. (I am irrationally annoyed by the segment in Jigsaw where the actions you need to perform are blatantly obvious, but only if you’re familiar with the works of the 20th-century French author Marcel Proust. You’re given some objects that appear in his books and need to take the actions that the character in the story took.)

To close out, a few of my favorite types of puzzles that tend toward the light bulb side:

  • There’s a passage that doesn’t allow you to bring object X through, but you need object X on the other side. I think it was zarf who said this was his favorite kind of classic text adventure puzzle, since it requires you to rethink your objects or rethink the map.
  • There’s something (like a revolving door) that switches parity every time you take a certain action (like walk through it), and you need to find a way to break that parity.
  • There’s an object that you used for its obvious purpose earlier in the game, but now you need to reuse it in a different way. The player is certain to know about the object already; they just have to connect it to their new problem.
  • Early in the game, you see something that you have no way to interact with (like an aluminum wall), so it seems to just be decorative. Later, you gain a way of interacting with that thing (like an aluminum-cutting tool that can slice through any aluminum), so the earlier object takes on new meaning.
15 Likes

I’ve always appreciated this quote from The Craft of the Adventure:

In an interactive medium, the beliefs and abilities of the protagonist are more than simply a painted backcloth, because the player participates in them. These special abilities might be called the ‘‘magic’’ in the game’s model world, in the broadest sense:

For magic and applied science alike the problem is how to subdue reality to
the wishes of man.
(C. S. Lewis, The Abolition of Man.)

In ‘The Witness’, for instance, the magic might be said to be the detective’s ability to arrest or call for forensic analysis, and in ‘Ruins’ we have the camera and the packing case. The magic is the imaginary fabric of the world, and it is as essential for the magic to have a coherent rationale as it is for the map to suggest a coherent geography.

The best puzzles for me participate in the magic of the world. “Magic” need not be actual magic, or even special abilities, but I think it’s best when the world operates by rules that are consistent and dependable. It’s great if they’re interesting, too, if that’s possible. My favorite puzzles are aligned with the world’s magic, and reliable deductions can be made because of that alignment.

This isn’t exactly the same as mimesis, which used to get talked about a lot, but rather more about how puzzles and solutions are at one with the characteristics of a world mechanically.

This could be something like the translation puzzles in Infidel or the way that in Zork III the player has to figure out that they need to become the dungeon master. Or a game with a cat that progresses by meowing at things. Etc.

Some games don’t do this at all! I just wrote a bit about Hollywood Hijinx, and it is mostly a funny place to put puzzles. It’s consistent in that way! But I will always prefer a bit of magic.

10 Likes

Great write up, I got three inspirations already out of your closing list

1 Like

IFWiki has a collection of links to articles about IF craft, including puzzles.

2 Likes

To add on to this - I’m a member of EJ’s puzzle hunt team IRL, and in addition to the hammers and lightbulbs we have one final guiding principle in puzzle creation: Is it fun?

That seems obvious, but it can be very easy to get wrapped around the axle designing clever and varied puzzles and forget about the fun entirely. The end result is often then something that is mechanically interesting on paper but tedious or frustrating to solve in practice. Luckily, it’s easy to catch that kind of thing in playtesting. If you keep an open mind about feedback and design your puzzles up front to have some amount of modifiable difficulty (if it’s easy to increase or decrease the number of steps involved, for example, or to control the availability of information) you can usually dial things in to a point that’ll make most people happy.

12 Likes

I’ll be curious to see what people think of some of them. I posted looking for testers but think I jumped the gun there as it’s not yet complete-able (plus it uses AI NPC responses, and while you can toggle it off, I still need to build out the static dialogs). I’ll get there, though!

1 Like