Despite the challenges of testing puzzle games (such as people only being able to play the game once), there are also good news: they generally don’t need quite as many playtests a “traditional” board game because there aren’t as many variables. This is not to say that you should skimp on playtests - it usually shows if a game hasn’t been tested enough - but, unless the game has branching paths, there is usually only one “correct” path, and that is to solve the puzzles. Even if players deviate from the obvious solving path, they can only go so far, so you don’t face the challenges that come with drawing cards, rolling dice, or even allowing players to make strategic decisions that change the course of the game. So that said, what are you testing for?
Solvability
The minimum criteria for a functional game (or individual puzzle) is solvability. Unless an individual puzzle is intertwined with the overall game, I usually create a rough draft of the mechanic alone, and ask a couple people to solve it. (Tip: I normally ask people who I know won’t be playtesting the full game to do these, so as not to spoil the puzzle for later.) If a couple people can solve it with no issue, then at least you already know you have a solvable puzzle. If not, fix any errors and keep making adjustments until they can. That might mean physical alignment of pieces, clarity or wording of a clue, presentation, layout, etc. After I know a puzzle is solvable, I mock up a draft of what a “final” version would look like. If it requires detailed or time consuming art then I just create a better looking draft that’s close to the finished look but won’t take too long to execute, as I still expect changes to be made. Then I test this second version in context with the narrative and the rest of the puzzles.
Details matter
It is possible to achieve clarity and precision, so don’t ignore the details. If something is meant to line up, keep adjusting it until it does. If two shapes look too similar, change one to avoid confusion. If you have similar colors on things that don’t go together and people keep trying to combine those components, change one of the colors. Some people try not to be too nit picky in their feedback, so it’s always beneficial to carefully observe how people interact with and interpret the clues. Be proactive about changing ambiguities. Finally, once the game has passed all these tests to your satisfaction, don’t forget that it must be tested with the final art. Puzzles can be made or broken by the images used. A change of one element could cloud up an otherwise logical puzzle, so make sure you do some final testing before sending it out into the world.
Take it or leave it?
Different people will give you contradictory feedback, and you will most certainly receive feedback and suggestions you disagree with, so how do you decide which advice to follow?
First of all, if multiple people are identifying the same problem, it’s something that needs to be fixed, so ignore them at your own peril. More often than not it’s just a matter of clarity or lack of cluing (a puzzle is too hard, doesn’t give enough direction); see my previous post about balancing difficulty.
I would say don’t worry if you think adding clues makes it less “challenging” - challenge shouldn’t come at the cost of fun. After all, you’re designing a game, not an exam. Same goes for puzzles that no one can solve without hints - they need more signposting. People like solving puzzles because it makes them feel smart, so allow them to have that satisfaction. My personal rubric is that at bare minimum 1 out of 4 people should be able to solve a single puzzle without a hint, though it’s worth noting that I usually try to have a variety of difficulties within a set of puzzles, so 1/4 would be on the hard end.
So what if just a couple people encounter a problem with a certain aspect? You can take this on a case by case basis and ask yourself: Is it something you noticed before they pointed it out? How difficult would it be to change? Would changing it improve the game? To what degree? Put yourself in the player’s shoes as you make this decision - what experience would you rather have?
Test your hint system
Lastly, don’t forget to test your hint system! After the first test of The Tale of Ord, I decided to set up my hints right from the start, so testers can experience the game the same way a regular would. This is important because there are even things about hints that might not be clear to people, and it’s better to catch those things early on. If you get stuck on a puzzle and the hints don’t make enough sense to help you move forward, that should be addressed. See my previous post on hint systems for more.
How do you make the most of your playtests? How do you decide which feedback to move forward with? Any other playtesting advice?