Due to illness and an unhealthy obsession in making rendered water look nice this post is a little late. Hopefully no harm has been caused :)
Figuring out a good puzzle is often a hard and tricky process. Sometimes a puzzles presents itself from story and environment naturally, but more often it is put in just to add some gameplay and/or slow the player down. This means first coming up with some kind of obstacle and then designing some sort of solution for overcoming it. During this process, and especially when "forcing" a puzzles into the game, one has to consider a couple of things. The most important of these are:
How fun a puzzle is to solve and how unique is it both determine the level of enjoyment a player gets from trying to figure and actually solving a puzzle. Solving the same kind of puzzles over and over is never fun and any appearance of a sliding puzzle is bound to bring forward feelings of unhappiness.
While a bit related to Enjoyment, a clever puzzles does not really need to be fun, it just needs to have a solution that makes the player think out of the box. A clever puzzles often includes using something in a non-obvious way and/or piecing together several fragments of information. If a game mainly relies on solving puzzles (such as Professor Layton), having clever solutions becomes extra important.
This means how well the puzzle fits with the story/world and is often the hardest part to accomplish. A puzzle with good world coherence adds realism and immersion to the game, while a bad one pulls the player out of the experience. In order to obtain high coherence a puzzle must fit with the story and also suite the world and not feel out-of-place.
When designing puzzles for Penumbra and our upcoming game, it is always a balance between these. Sometimes a puzzle might fit perfectly with the story but just be really dull and sometimes a fun puzzle does not fit at all with the game world. It is almost impossible to come up with a puzzle that "score" high in all the above criteria, so one has to concentrate on something.
In horror games, where immersion is key, it is probably best to always make sure that no puzzle feels out of place. For some reason many action based horror games seem to forget this and are filled with mood breaking puzzles. In Penumbra we did our best to have as high world-coherence and often had to sacrifice other criteria in order to do so. This is one of the reasons why Requiem contains so little story, we wanted for once to concentrate on the making fun and clever puzzles.
Most (at least we hope so!) puzzles in Penumbra where not dull and stupid, but often we concentrated on either making it clever or fun. In the cryo-chamber, getting the head out of the jar was not a very rewarding puzzles to solve, but the main idea there was to let the player do something fun (who does not like playing around with severed head?), instead of teasing the player's brain. Figuring out how to enter the cryo-chamber was instead an attempt at making a clever puzzles and required several pieces of information to be linked.
By trying to vary puzzles like this we hope to have made the experience more interesting. As discussed earlier, games does not need focus on creating joyful feelings all the time. By letting the player sweat over a more complicated task, more emotions can be added to the game and end up being a more rewarding experience. Adding instances of more fun and simple puzzle in between breaks things up and make the brain-teasing parts stand out more.
This brings me to the final issues one has to consider. Difficulty. I did not include it in the list above because, while an important thing to ponder, it is quite a different beast. The main problem lies in that when a solution is known it is no longer hard to solve, and thus it can be hard for designer to now the difficulty of a problem. Even so, it is a very important part of the gameplay and the time spent pondering a puzzle plays a large role in the gameplay flow. At times it might be fitting to throw a harder challenge at the player and other times the player should be able to solve it quickly. Especially when a situation is meant to be frightening, having the player scribbling on a note in the "real world" is not good for the mood.
Pretty much the only thing that can be used to test difficulty is extensive play testing, but this being time consuming and expensive (especially since the same person can not reliably test something twice) other methods are needed. I usually try to "wipe" my mind, think myself in the situation of the first time player and imagine the moves she would make. This is actually not far from the tactic used when designing scary situations, something that also greatly relies on an unknowing player. Designing puzzles is actually kind of related to creating a horror atmosphere in that one has to try mess with another person's mind and supply hints, confuse, etc in order to create a satisfying experience. This is yet another reasons why horror games and puzzles are such a good fit.
What do you think is most important criteria for a good puzzles? As always we are also eager to hear feedback on puzzles present in Penumbra with the above in mind!