What makes a 'best puzzle'?

What makes a ‘Best Puzzle’? (Edit: this is a repost of a several-years-old article)

Every year, public voters award the XYZZY Best Individual Puzzle award to one puzzle in one game. These are the puzzles that most captured the community’s fancy in some way, whether through complexity, inventiveness, or even emotion. Of all the XYZZY, the Best Individual Puzzle award seemed at first to me the most like a grab-bag of random games, with little connection between different winners.

However, I was determined to find out what, if anything, they had in common. After playing carefully through all the winning puzzles, I was surprised to find that they were remarkably similar to each other, falling into a few broad categories. Avoiding spoilers as much as possible, I’d like to to discuss these. The main categories are Learning a System, Iterative or Babel Fish type Puzzles, and Sudden Insight. These are not rigid; the Babel Fish type puzzles, for instance, are essentially a blend between the other two categories.

=====Learning a System======

Roughly half of the best puzzles presented a complex system, sometimes taking up the entire game, which you have to learn how to manipulate through experimentation. These puzzles are hard to spoil, because it’s clear from the outset what you have to do and what tools you need; the main difficulty is figuring out which tools do what.

The first such Best Puzzle is the language puzzle in The Edifice. The Edifice was the 1997 IFComp winner, and had the player visit three or four scenarios in the evolution of man, from the creation of tools and fire to language. In the language puzzle, the player meets an NPC who communicates entirely in a language of the author’s invention. By experimenting with the words you hear, you gradually become able to communicate with the NPC and ask them specific questions.

A language puzzle appeared again in 2001 with The Gostak, a highly polarizing game that won the Golden Banana of Discord in the IFComp (an award for having the highest standard deviation in votes). This game is written completely in a sort of pseudo english, based on an old artificial sentence that says ‘The Gostak distims the doshes’. Just as in the Edifice, the player must experiment with the new words and their responses to determine how to play the game.

The Best Puzzles in 2004 and 2014 involved time travel puzzles. Both of these games (All Things Devours and Fifteen Minutes) are one-puzzle games with essentially the same premise: you have run out of time, you have access to a time machine, and you’re going to need more than one version of yourself to get the job done. The two games differ in how they handle the versions of yourself: All Things Devours has a large map, and the difficulty in the game is learning how to avoid your copies to prevent a paradox from occurring; while Fifteen Minutes is a one-room game that relishes having all of the copies present at once and interacting. Both games are best approached with copious note-taking materials.

The last two System-Based Best Puzzles are found in the games Delightful Wallpaper and An Act of Murder. The first game, by Andrew Plotkin, features a player who cannot manipulate anything directly, and who can only move about. The house they are in, however, reacts to their movement, with movement in one part of the house triggering doors and gates to move on the other side of the house. The game provides a helpful note-taking system, but it is fiendishly difficult to map out a path that will open the doors you need. The second game, An Act of Murder, is a randomized murder mystery that again requires careful note-taking as you establish alibis and determine motive.

What do these system games have in common? For one, they don’t overwhelm the player. Other attempts at language puzzles, for instance, have dumped big dictionaries or grammars on the player. The Gostak and the Edifice, meanwhile, present a sentence or a paragraph, and let you go from there. Other time travel puzzles (of which there are many) may force you to interact with copies of yourself too early.

Conversely, these Best Puzzles also grow complex enough for the player to feel like they really learned something when they succeed. The Gostak requires a fairly complicated sequence of actions to win, and the goal in the Edifice is to communicate a pretty complex request. The maze in Delightful Wallpaper was quite large, and the time travel puzzle in Fifteen Minutes has you interacting with half a dozen or more copies of yourself.

Finally, these puzzles tend to be clear about the function of each new piece. Delightful Wallpaper will actually write down for you what you’ve learned at each stage, making it completely unambiguous. The use of each new word in the Gostak or the Edifice is made clear by the many situations it appears in, and so on.

So it appears that one type of puzzle that people regard as great is a complicated system where one is clearly taught from basic principles to a complicated finale. This is the same as learning a system in real life, from mathematics to physics to cooking to speaking a language. These puzzles train you in a (in-game) useful skill, and they train you well.

====Iterative or Babel Fish Puzzles====

The term Babel Fish puzzle comes from the Infocom game HItchhiker’s Guide to the Galaxy, where the player tried to get a babel fish from a vending machine. Each attempt goes hilariously wrong, with the fish falling in a hole or getting vacuumed up, and as the player tries to fix each problem, a new problem arises. The term Babel Fish Puzzle is now used to describe a puzzle with an iterative approach, where each new try yields new information.

The puzzles that I include in this category are disconnecting the internet in Violet and the whole games of Rematch and Lock and Key. Some of these could be classified as Learning the System or Sudden Intuition games, but I feel they belong here.

It is easy to spoil the puzzle in Violet, but much of its charm lies in the writing more than the puzzle structure. Each step makes you laugh at your own lack of self-control. The final solution involves some actions that may be counterintuitive and mildly color your perception of the PC for the rest of the game.

In Rematch and Lock and Key, you set up a system, see how it works, and restart. Rematch is largely about getting a single complicated command correct; the author warns you that the final command is so long that the parser had to be partially rewritten to accept it, and the gameplay consists of iteratively experimenting with lengthier commands until the correct one is found. Lock and Key has you designing a dungeon with a selection of about 16 pre-made rooms which you must make into a maze, before watching an adventurer break through them. Like Rematch, it’s clear what each part of the puzzle does, it’s just assembling them in the right order and the right length that’s difficult. Rematch keeps the puzzle interesting with a super-dramatic recurring event, and Lock and Key keeps it interesting with hilarious descriptions.

Interestingly, the key to success in these puzzles seems to be the writing and atmosphere more than the puzzle design itself. The frequent failures that are shoved in your face are only made palatable by the amusing or compelling response that you get in return. A Babel fish puzzle without good writing would only be an exercise in frustration.

====Sudden Intuition Puzzles====

These puzzles are essentially impossible to discuss without spoiling, so I’ll just talk about all of them together in general terms. These are puzzles where you are presented with a confusing or frustrating situation, and you must have a sudden flash of insight to understand the puzzle.

So many authors have tried to write compelling Sudden Intuition puzzles and failed. In looking at the Best Individual Puzzle winners, it seems clear that there are a few ingredients to success:

  1. The flash of intuition occurs in a constrained environment. In each of these games, the flash of intuition occurs when almost all other puzzles are out of play, where the player is in a small area and can only work on the one puzzle, and where it is clear that you need to do something special to move onward.
  2. The flash of intuition involves some action that you’ve been taught before. In every case except perhaps one, there was an earlier puzzle that you solved by performing the exact same action, just for a different purpose. This trains the player to know that your intended solution is actually possible.
  3. The puzzle involves using something familiar in an unfamiliar way. This is what makes the puzzle interesting.
  4. There is strong dramatic tension pushing you forward, making it vital to solve the puzzle (often the threat of death).

Since I don’t want to spoil any actual examples, I’ll use an example from literature to illustrate this. In The Fellowship of the Ring, Gandalf and the fellowship are stuck outside of the walls of Moria, trying to get in, with enemies and evil mountains behind. They are in a narrow, dead-end valley with a lake on one side and the gates on another. The gates of Moria say ‘Speak, friend, and enter’. Gandalf interprets this as the door needing a magical password. With nothing to do but think, the fellowship spends hours working on the solution as Gandalf speaks word after word of power.

Finally, with the help of the hobbits, Gandalf laughs and realizes that the correct translation was ‘Speak Friend and enter’. He says the word ‘Friend’ and they go in, just as a monster attacks.

Note that the characters were placed in a constrained situation with no other distractions. They have a compelling story-based need to go in, with their very lives being threatened. The solution was unconventional; the classic idea of a ‘password’ was turned on its head, as being a word that everyone should know instead of a word that no-one should know. And there were, in fact, hints in the book. Earlier in the story, much is made of the former friendship of dwarves and elves; when the solution is discovered, it is readily apparent that it fits with what we know of the old days.

====The last puzzle: Four Hats====

The last Best Puzzle is different from all of the rest. In 2011, four IFComp games featured a recurring character who was looking for his hat. Most people did not notice or chalked it up to coincidence (like the year that several games featured a giant squid). The four games were quite different: Cold Iron (by an anonymous Plotkin) was a short, mostly puzzle-less story about a world where faeries may or may not exist; Playing Games was a sequence of three mazes made in ASCII art with a threadbare story tying it together; the Last Day of Summer was a short atmospheric piece about regrets in a village; and the Life and Deaths of Doctor M was a full-blown journey through the afterlife, exploring the questionable actions of a doctor through flashbacks.

However, all had a general setting of late-1800’s to early 1900’s as shown in architecture and clothing (for Doctor M, in the afterlife section only), and a feeling of remembrance and wistfulness. And, of course, the recurring theme of the hat.

The puzzle consisted of transferring passwords between the game in various forms. I’ll avoid spoilers, but the idea is that one game might have, say, an open combination lock with the combination visible, and the next game has the same combination lock, but closed. Or one game might have a society where no one talks to you unless you bow three times, and the next game has the same society, but doesn’t tell you to bow. So only by playing all 4 do you see the whole puzzle.

I believe that this puzzle won for originality and cleverness, mostly, as it was a bit unfair in practice: no one was even able to deduce its existence, let alone solve it. However, it does have some things in common with the other Best Puzzles. For instance, there is a real sense of learning and mastery as you proceed from game to game, because you have knowledge that no one else in the game world does. And you do use items in unfamiliar ways, providing a new insight to previously mundane materials. In any case, this is certainly the most unusual of the Best Individual Puzzles.

====Conclusion====

While the Best Individual Puzzles are a mixed bag, there are some consistent trends in what voters like. Besides the hat puzzle, all of the puzzles are fair; either the rules are clearly laid down, or the puzzle is set up in a way that more than half of the players will stumble onto the solution. All of the puzzles teach the player something, whether a new system or a new way of looking at the game world. And, which I have not mentioned before, each of the puzzles was placed in a game that was polished, bug-free, and frequently well-written.

What implications does this have for authors? It’s hard to come up with specific recommendations, but one thing that many authors seem to neglect is fairness. The puzzles in other games that provide misleading error messages, or provide huge info dumps all at once, or which are purposely obfuscated, or which neglect to give any hint as to your purpose, or which require you to ‘guess the verb’, are almost uncountable in number. The results of this survey of puzzles indicate that just as much craft should go into leading the player into the solution as went into creating it in the first place.

24 Likes

Also relevant: Emily Short’s article on the 2013 XYZZY Best Individual Puzzle category, identifying categories: extent, explorability, surprise, ingenuity, originality, fairness/accessibility, structural integration, and narrative integration.

https://xyzzyawards.org/?p=386

9 Likes

I think there are (at least for me) two kinds of “best puzzle.”. The first kind are the ones that provoke a Eureka! moment and tend to be non linear i.e. the T.S. Eliot Hollow Men boatman solution from Curses. The type that have you rising from your pillow at 2 am to try them.

The second type are the Heath Robinson conundrums i.e. the Babel Fish from HHG, the guard dog from Not Just An Ordinary Ballerina and my personal favourite, the gold key from Castle Ralf. The type requiring old fashioned paper and ink stained lips.

I think the latter problem can outdo Nelson / Partington / Meretzky for sheer classic construction yet it remains, like Finding Martin, obstinately obscure.

2 Likes