Sander Vanhove's Blog

Puzzle games are the biggest mobile game genre in the UK and US and reach one of the widest demographics. Everybody loves a good puzzle. But what are puzzles and how do we design one?

What is a "puzzle"?

A puzzle is a very strange kind of game. You can not play a puzzle against somebody else (except maybe if you time it). You can not replay a puzzle, once solved replaying the puzzle is just a chore.

I think Jesse Schell put it best in his book The Art of Game Design:

A puzzle is a game with a dominant strategy.

A dominant strategy means that there is a strategy that always makes you win the game. Furthermore, the objective of a puzzle is finding that dominant strategy. This sounds about right, once you know how to solve a puzzle (find the dominant strategy) you will always be able to solve it. But I think this definition misses one component.

All state should be visible

All the state of the puzzle at hand should be visible. Else a player is will not be able to think logically about the individual components and work their way towards a solution.

Showing all state makes everything predictable and enables the player to reason about outcomes of certain moves or actions.

Of course there could be multiple stages to solving a puzzle, where information is revealed slowly, but all information for each separate step should be present. In this way, each step is a separate puzzle.

Let's say some state is hidden or unpredictable. This could be in the form of cards that lay face down on the table or a roll of the dice. All of a sudden the player has to take this in to account too and will have to plan around these unknowns. Although these games could look like puzzles, I would lean more toward calling them strategy games because you have to strategize around unknowns, doesn't matter if these originate from the game itself or an opponent. This is why a puzzle inherently can't be a competition, influencing another player's board state means introducing randomness in the puzzle and consequently that player is not able to think ahead logically but has to think probabilistically.

Ideally you can solve a puzzle without performing any actions. Purely by observing the starting state of a puzzle, you should be able to solve the challenge without having to actually go through the motions of doing so. You can solve a sudoku without writing down any numbers or a Monument Valley level without moving even one step. Of course this is not how we solve puzzles, we solve them step by step and only solve the sub-problems directly at hand, but this is more to blame on our limited brain capacity and can be overcome by training.

This is why I would not call Candy Crush or Tetris puzzle games for example. Not all state is perceivable. Like said earlier, in these games you try to strategize around unknowns by keeping the board open for anything the game throws at you for example.

That doesn't mean that puzzles don't involve skill, Angry Birds for example is a puzzle that involves some skill of aiming the bird. Catapulting a bird in exactly the same angle with exactly the same velocity however will give you exactly the same result every time.

Let's put this all in to one sentence:

A puzzle is a game with a dominant strategy where all state is visible.

Some tips for designing puzzles

Ok, we defined what a puzzle is, now let's go over some tips for designing good puzzles.

1. Use a small amount of mechanics

The best puzzles have very tight core of mechanics.

Think of Sudoku:

• Write the numbers from 1 to 9 in squares.
• Never put the same number twice in one row.
• Never put the same number twice in one column.
• Never put the same number twice in one of the marked 3*3 groups (blocks) of squares.

Or Crossword Puzzles:

• Put the letters of the described word in the designated squares.

Or one of my own games Silent Dot:

• Get the dot to the triangle.
• You can move from hexagonal tile to hexagonal tile.
• Some tiles aren't available to move to depending on a pattern that is visible.

Interesting and captivating puzzles arise from a feeling of understanding what needs to happen, but being astounded that it is harder than it seems.

When creating a puzzle try to cut as many fluff-mechanics as possible.

2. Show clear progress

Make sure the player gets a feeling of advancing towards their goal while solving a puzzle. You don't want to be stumbling around in the dark.

For sudokus and crossword puzzles this is easy: every square that is filled with a number or letter is one square closer to solving the puzzle.

I'm a little afraid my puzzle game, Silent Dot, is not the best in this department. There is no good metric of defining that you are getting closer to solving the puzzle. Even the distance in tiles from the dot to the end tile is not a good measure because it could be that the pattern of one of the tiles doesn't lineup and you have to start over again.

3. Incorporate a perceptual shift

A perceptual shift is the "Aha!" moment of a puzzle. It teaches a new point of view the player hadn't considered yet. This kind of moment only occurs after you establish how the puzzle should work in an obvious way. From the moment the player thinks they found a dominant strategy they can apply to easily solve every puzzle, you hit them with a complete novel way of thinking.

In essence you are teaching your player something new. First they will have to do the perceptual shift themselves and afterward they need to recognize when to use which point of view.

To pull this of you need to build up to it. Start out with some easy levels and slowly make them harder, up until a point that the players current view on the problem is not sufficient enough and they need to find a new way of perceiving it. In sudokus for example, a perceptual shift happens when you know that a number will certainly be placed in one row of a block, you can rule this number out within that row without knowing exactly where it will be placed.

For example in the next snip-it of a sudoku, we see that, without knowing exactly where the "1" will be in the middle block, we can still deduce where the 1 will be placed in the left block.

The problem with a perceptual shift is that when players don't see it, they will not get it and just feel frustrated not being able to solve your puzzle. But when done right, a perceptual shift will result in the most satisfying solve and will make your player much better equipped to solve future puzzles.

Conclusion

Puzzles are a very interesting type of game. We defined what a puzzle should be and looked at some tips for creating our own. Now go forth, work on your own puzzle concepts and please send them to me!