4. We Made a Game That Was Too Hard to "Get"
There aren't many games like SpaceChem, which makes it spectacularly hard to explain, especially to people who are only familiar with more "mainstream" titles. To make matters worse, the surface appearance of the game -- fake chemistry, often mistaken for real chemistry -- completely belies the game's addictive Portal-like problem solving. To complicate things further, our presentation of fake chemistry lacks the appeal (i.e. sexy scientists) to draw curious viewers in! The end result is a game that is nearly impossible to discover, try, and buy.
Fortunately for us, SpaceChem is extremely addictive! And, as is the case with most addictive substances, it is the responsibility of your questionable friends to get you hooked. Partially because of our planned "community features" and partially because it's just what great fans do, SpaceChem's early adopters took the time to tell their friends that they had to try it.
Although I doubt anyone was able to explain SpaceChem and make it sound like fun ("You went to university for computer science, right? Well, I have the game for you!"), making a game that our fans demanded their friends try got us pretty far.
5. We Never Got the Tutorial Right
Historically, my love for making complicated games has outpaced my ability to explain them properly. SpaceChem is no exception to this trend.
When we launched, the tutorial for SpaceChem consisted of 12 puzzles that covered the basics -- inputs, outputs, bonding, unbonding, reactors, and pipelines. Two of the 12 puzzles were "walkthroughs", which included explicit instructions for building a solution and explained, step by step, what each piece was responsible for. Additionally, there are also 13 "info screens" spread throughout the entire game explaining new, high-level concepts (such as atoms + bonds = molecules, and what to do on defense missions) as they are encountered.
The responses of players who played through the tutorials ranged from immediate understanding to complete incomprehension (something that I think is actually quite uncommon for modern games, given the combination of better design practices and better "game literacy" among players). Some of the reasons I suspect for this are:
We failed to clearly show the objective of a puzzle. The primary goal of a SpaceChem puzzle is to take "inputs" and convert them into the specified required "outputs". For anyone in the software industry this is an obvious endeavor, but for the general population it's not a given, especially in the context of fake chemistry.
Although the first puzzle (a "walkthrough") clearly demonstrated inputs and outputs, most players follow the instructions without recognizing the purpose of what they're doing. We partially resolved this by including a video tutorial that focuses on the high-level goal of SpaceChem -- transforming inputs to outputs -- giving the player some context when they reach the first walkthrough.
We failed to make the game start simply enough. To build a level in SpaceChem, it's necessary to build an entire "loop" with the minimum following components: input, grab, arrows, drop, and output. Considering most players barely understood their goal, forcing a minimum of five different tasks on them from the beginning is overwhelming. This problem could have been fixed by changing the design space to have a smaller "minimum possible solution", although it wasn't much of an option post-release.
We exposed too many details too quickly. SpaceChem is filled with lots of details and rules, such as what can be moved, what can be predicted, and what is required to solve a puzzle. Although the interaction of these rules makes SpaceChem open-ended and emergent, it also makes it confusing, especially when players misinterpret what they see.
A common example of this concerns "waldos", SpaceChem's programmable manipulators. The default configuration for a reactor consists of a red waldo at the top and a blue waldo at the bottom; because this is the default setup, many players assume that the red waldo can only be used on the top and the blue waldo can only be used on the bottom. We mitigated this and similar cases by tweaking the tutorial puzzles to include counter-examples, although the problem is by no means fixed.
We used too much text to explain things. There were many situations where we were forced to use text to explain rules and nuances of the game. In retrospect, this is a clear sign that we needed to change the game, not explain it more forcefully.
You lost me when you started talking about molecules.
All things considered, the development and release of SpaceChem went much better than anyone on the team anticipated. I had originally hoped to cover our costs and make a little money, but ended up leaving my job and starting a game studio to work on new titles. New titles, I might add, that are hopefully more accessible than SpaceChem!