When a good game designer is not a good player.
Let’s put a thing aside from the very beginning:
Every game out there is a problem solving experience, so puzzles are involve as a fundamental part of any game in existence. I wrote a post about that point a few weeks ago. On the other hand, it is common practice among game designers to create problems that can be solved in a number of different ways. When you play football there is a lot of possibilities at hand to score. Other games aren’t that open, but still offer a fair amount of freedom to the player.
The idea behind that is to avoid First Order Optimal Strategies (FOOS). Here there something about FOO’s (and related subjects) I posted months ago, but summarizing a FOOS is a strategy that represents the optimal way to solve the problem the game has to offer. If a FOOS exists, there is little the players can do to play the game in a different way, or in other words: as the problem has been solved, the game doesn’t have anything more to offer.
An excellent example is Tic-Tac-Toe as described by Raph Koster in his book A Theory of Fun. When you realize that any starting position have a very definitive answer in order to win, the game have nothing to offer from that very moment.
When there is no FOOS whatsoever, the players can choose a way that seems (to them) as the optimal way to play. The consequence of that fact is called Emergent Gameplay. Emergent Gameplay are all the solutions players can give to a problem in a game, even (and specially) involving those not considered by the designers. Football, again, is a game with a lot of degrees of emergent gameplay. There is a whole century of strategies in football. Emergent gameplay is a desirable characteristic in a game.
Other desirable characteristic in a game is Replayability, which could be considered as a consequence of emergent gameplay, but in fact there is a loop involved: fair amount of freedom for emergent gameplay means a fair amount of replayability, and the other way around. You can think in both as a separate, but related, entities when designing a game.
But when you look closely, an special case of the last definitions arises: puzzles. As Jesse Schell points out in The Art of Game Design, there is (and will always be) a discussion about whether puzzles are or not in fact games. Puzzles have an answer, that means there is a FOOS involved. Therefore, puzzles can’t be replayed (at least an instance of a type of puzzle) nor give to the player emergent gameplay at all. What’s the point of puzzle then? Schell answers: Puzzles are games where the whole point is to find the FOOS (but he calls them Dominant Strategies).
Let’s focus our attention in Crosswords, a well known type of puzzles. As said, crosswords have an answer, there is no replayability in every instance of a crossword. That could mean players that solve crosswords have developed some sort of strategy in order to do so. An skill is created (or developed) and refined until there is little (or no) challenge. You can derive the same reasoning for other puzzles.
Well, that statement seems pretty obviuos, but brings us to the core of this post: what about puzzle making? are those skills needed in order to solve puzzles the same needed to create that puzzle? the answer is no. As simple as that. In my opinion, there is a little more in puzzle making that in puzzle solving.
(Of course, you can always see “puzzle making” as a problem to be solved, therefore transforming it in a puzzle solving experience, but that is not the subject of this post).
See, I picked crosswords as an example not only because they are well know, but because in this era of Google, Wikipedia and Wolfram Alpha, crosswords look outdated. A lot of people compare crosswords to Sudokus and say: Sudokus are superior because you can’t answer them by using Google (which is false, but I know you get the point). As viewed only as a memory training hobby, crosswords do look outdated, but that’s no the case.
What are the skills needed to solve a crossword? memory is one, that’s ok, but you can’t pretend to solve every single crossword based on memory (unless is a very bad designed set of crossword). Besides memory, there is a contextual inference skill involved: you can totally discover (or remember) a word based on the context, even without reading the hint given by the puzzle. Contextual inference is not a skill fueled by Google searching.
But there is (at least) another skill involved.
Ask yourself, just basing your judgement in the number of letters of a particular word, what is the best strategy at hand: to try to discover (or remember) a word with a lot of letters, thus rising the chance of completing a lot of others words in consequence; or try to complete a lot of short words that can be discovered (or remembered) fast and then use the contextual inference skill to discover (or remember) the words with the biggest number of letters.
Is not and easy answer. When you choose to look up a definition in Google you are choosing it pretty much based in answering the above question, and that choice is not fueled by Google searching either.
When designing crosswords (and I designed a few in order to write this post) it seems to me that there is no particular way to balance what I call bottom-up/above-down strategies (BU-AD). The crossword itself, as a puzzle with a FOOS involved, guides you to a particular way to design it (I insist, at least in my very short experience prior to write this post).
Crosswords, as seen by know, don’t look so outdated.
You can do a similar analysis with Sudokus or any other puzzle out there.
As a designer, your work consist in create an instance of a puzzle that involves some little amount of thinking that leads to discover the FOOS of it. Designing for contextual inference and BU-AD strategies is a way to preserve some sort of challenge. When a player is presented with a crossword, you can totally design one that involved the conscious and rational use of those skills, even when they mean the use of Google in the process.
That skill is not listed for puzzle solving, but for puzzle making. Leading the way to a particular experience is a very special skill required for game designing (among others I wont discuss here), that actually in the case of puzzle is the exact opposite of what players do while playing.
And keep in mind that contextual inference and BU-AD strategies aren’t exclusive for crosswords. A lot of instances of those skills can be named. To bring up just two, contextual inference is used for environmental storytelling (see Bioshock and Half Life 2 level design) and BU-AD are the kind of things you use while choosing which quest to complete in a RPG.
Dark Recon and Orbita involve a lot of puzzle making- solving, so in the next posts I will take it from here and write about how we are creating those puzzles and what kind of things we derived as game design kwnoledge, that can be useful for other designers/developers.