Just Ice postmortem

I found the scale and magnitude of MiniJam 45 very comfortable, so I preemptively signed up for the next one.  The theme was JUSTICE. I, like many of the other participants, thought “Oh ho ho! Justice can be split into two words, JUST and ICE!  How droll!” I mostly don’t want to make games that don’t have wizards in them, and ICE felt more like a wizard thing than, like, jurisprudence.

Unlike the MiniJam 45 game, I spent a week or so percolating on it and had a pretty good idea of what I wanted to do before the additional constraint was announced.  Luckily, it was “the game can only use 5 colors,” and I don’t find that having a limited color palette cramps my style in the slightest.

I set out to make a clicker game with finite scope and a little bit of a resource management / optimization puzzle built into it.  I always liked the part of Civilization 1 where you build extra walls and towers onto your castle as a reward for progress, and I wanted to evoke that.  I guess it ended up being more like the town management parts of the Heroes of Might and Magic games.

My other goal was to become more comfortable with Javascript, which is a desire that constantly gnaws at me like an unfinished homework assignment.  It’s just similar enough to languages I’m comfortable with to be extremely frustrating when stuff doesn’t work. Online documentation for it spans 20+ years and a million different frameworks for turning a basic web page into a 40-megabyte trash monster.

Anyway.  Over the last couple of years I spent some time messing around with Dwitter (http://dwitter.net) so I had some basic facility with getting graphics onto a CANVAS element.  It worked out okay -- I think I’ll try using Javascript again the next time I’m doing a non-engine-specific jam.

I feel slightly weird about asking people to playtest these indulgent personal projects on such a short timeframe, and I knew I wasn’t going to be able to take the time to do enough of it myself, so I built it with automated testing in mind -- I wrote a routine to play through the game over and over again, making weighted random action choices, and logged the results.  I tweaked values until this resulted in a nice range of scores. I never saw the bot get a perfect score, so I shipped it without being entirely sure it was possible to score 100%. A guy on Twitter sent me a screenshot of having done it, though, so apparently it is!

I had to re-publish once because I’m incapable of spelling SORCERER correctly on the first try.  It’s just so much more intimidating and cool when it ends with OR!