Level design is difficult.

Orginally posted on RetroEpic's own blog.

We (here at RetroEpic Software) are just under two weeks from re-releasing our first commercial game A Day in the Woods to iOS and we’re still making changes to our puzzle levels. Why? Because level design is difficult. And level design for a puzzle game is a special kind of mind-bending madness that you’ll both love and hate at the same time.

I was first cast as level designer for the game back in 2011, when A Day in the Woods was heading towards its first release. At the time, I was pretty new to level design but hit the ground running with enthusiasm and self-surety.

And then I found out we were aiming for 60 levels and my stomach sank.

One of the first things I had to learn, and let my ego recover from discovering, was that in designing a level you need to consider multiple solutions. I had jumped into putting some levels together and was pretty proud of my efforts but I was soon to find that they weren’t fun to play. The first couple of puzzles I put together for A Day in the Woods has a single solution path that tended to be lengthy and dull. As a level designer, you need to figure out how to use all the elements of the game to put together a level that lets players explore multiple solutions without repeating the same problem solving task multiple times within the same puzzle. That is no mean feat and really only comes about by doing. Make levels. Have others play them. Make new levels. Rinse and repeat, and try not to let your ego get the best of you in the process. Speaking of testing…

The sooner you get the input of others on your levels, the better. People approach puzzles differently, because they’re all wired in weird ways. This means that even when you’ve built the puzzle yourself, you might not even be able to see the optimal solution. Getting others involved in the process as early as possible helps you get to know the mechanics even better, and will inspire better levels. Even though I started my first pass having a fair understanding of the main mechanics, there was so much more to it I'd been missing – creative uses and solutions! Even now, we’re discovering more optimal ways to solve puzzles that we’ve had in the game for years. (I have no doubt that there will be players who will find even better ways to solve them once the game is out there so, please, if you do find a shorter route to granny’s cottage pat yourself on back and then let us know!) Getting players to test your game is also just a generally important rule in game design. It helps you find problems and bugs - especially ones that help players “cheat” the levels and solve the puzzle without having to follow the rules that are supposed to be programmed in. Also, try not to think too many murderous thoughts when someone beats what you were absolutely sure was the best time, or least moves, for that level. It’s a huge compliment to have people spend enough time on your puzzles to discover new tricks!

Of course, you need to introduce your players to the mechanics so that they can best you. One big problem with working so closely with all the game mechanics and getting to know the game so intimately (as you do in designing the levels), is that it’s easy to forget what it’s like to be introduced to the game for the first time. It’s important to ensure that you’re creating a solid player progression that allows the player to meet each mechanic at an appropriate pace – without getting bored – to be able to have all the tools at hand when it comes to solving the more complex puzzles. Part of revitalising A Day in the Woods for iOS included a serious overhaul of the player progression. Some levels, we realised, were too complex to appear as early as they did in the game which would lead to player drop-off out of frustration. Some levels, on the other hand, suffered from a bit of grind and dullness. To solve this problem, we first graphed the appearance of mechanics and items in the game to see where the peaks and lulls were. To smooth out the curve, we swapped elements out, moved levels around and even re-created whole levels. Balancing the introduction of new items and mechanics without overwhelming the player with too much at once, or not giving them enough to play with early on through under-estimating their ability to grasp a new mechanic quickly, is a tricky thing indeed. See point 2 about testing, because this is pretty much the best way to approach this problem.

If you’re challenging the player to finish a puzzle in a certain amount of time or moves, you’re going to need a way to record the optimal solution. Early on in making A Day in the Woods, I wanted a tool that could auto-play a level so that the player could see what route he or she had taken and how to improve on his or her play-through. This replay system turned out to be a sanity saver because it also allowed us to record the level solutions and store them for future reference.

One interesting part of the process of designing levels for A Day in the Woods the first time round was that I was making the puzzles with whitebox assets while the artists were hard at work making beautiful pieces. Quite often, as the new assets arrived, I felt more inspired to set up puzzles differently because I wanted to show off each asset. Let your game inspire you. Never forget that your levels are part of a whole new world you and your team are creating. Don’t let your levels be designed separately from this world.

Finally, get your game out there! There’s no better feeling than having your game out there in the wild and players enjoying the levels you’ve meticulously put together. A Day in the Woods is set to release on 2 April 2015 on iOS.