How to make interesting mazes
Solving a maze is about making choices and you make those choices every time you find a branch. But sometimes not every branch is an interesting choice, and I’ll illustrate some of those cases that may affect the quality of a maze. By avoiding those cases you can make a more interesting maze.
I’ll focus only on classic orthogonal mazes but some of the cases highlighted in this article could be applied to other kinds of mazes. Also, some of these cases will not apply to 1st person mazes, like the ones in video games or hedge mazes because you have less information available while traversing the maze. And, obviously, making an interesting maze can also be a matter of graphic design and adding other interesting elements like bridges, but I’ll leave that outside. All the following mazes are supposed to be traveled from top-left to bottom-right.
Short dead ends
It’s too easy to spot a dead-end without traversing it if you make it too short. In the following example, in intersection A it’s easy to identify the wrong path without traversing it, but in intersection B you must choose one of the paths to identify which one is wrong. That means you don’t need to do an interesting choice in A because it’s too easy to see the correct path.
Try to avoid dead ends that are too short. Instead, lengthen them and add a few turns until it’s no longer obvious that it’s a dead end. When designing a maze you may end with small empty spaces and you could be tempted to fill them with short dead ends. Instead, try to fill them by extending another path.
Compact dead ends
Similar to the previous point, a compact dead-end is easier to spot without traversing it. Look at the following example:
Both paths have the same length, but the red one is easy to spot without going through it. For the green one is more probable that you choose it to see if it’s a dead end. Part of the issue, in this case, is related to the next point.
U shapes
The problem with U shapes is that you don’t follow the path as it was designed, you see a U shape and you go straight ahead as if there were no wall there. This is illustrated in the following examples:
A few U shapes to fill empty spaces when you’re designing a maze may be harmless, but if your maze has a lot of them, then it could mean you’re wasting space that could be used for more interesting paths.
Branches near the end
Creating branches near the end is useless because there will be a part of the maze that will never be explored. In the following example, there is a branch near the end that expands to more branches (highlighted in red), but it’s evident that the exit is just one step ahead, so no one will choose to visit that part of the maze.
I’m not taking into account the case of people that solve mazes starting from the exit, because anyway, that’s not the way a maze is supposed to be solved.
Too few branches
Solving a maze is about making choices, but if a maze doesn’t have enough branches, then there will be not enough decisions to make it interesting. The maze will have many long paths without branches and will be boring to follow those paths without taking any decision. And of course will be a boring maze to solve, like the following one:
Too many branches
On the other hand, if a maze has too many branches, that means there will be shorter dead ends and as we saw before, that means those dead ends will not be chosen, so it will lead to fewer decisions to take. In the following example, there are many paths that you’ll avoid because is too obvious they lead to nowhere:
In the end, you must find the right balance in the number of branches.
Final thoughts
Some of the points highlighted above are the reason most of the mazes generated by an algorithm are less interesting than the ones made by a human. Most of the algorithms don’t take into account considerations like the previous ones.
If you have more ideas about how to make a maze interesting, let me know in the comments.
