A success of scope: Why GGJ 2014 was my favourite jam yet

So - game jams are not like regular game development. They're not even like regular work that has a deadline. You go in blind. You may flounder before you even touch a keyboard. You often don't have time to go back and start again if your idea isn't working. Most of all, 48 hours (even without much sleep) isn't enough time to make a big RPG, or to get things COMPLETELY polished. You have to plan at an incredibly small scale in order to get anything at all across the line.

In several jams, I've planned for games that were too complex, that had too many moving parts - and then fell short of that plan. This Jam, GGJ2014, I went in with no ideas in my head ready to go, no big plans. I didn't even get time to listen to the keynote (although I went back later and watched it). I saw the phrase, listened to people pitch their ideas, and then got up and pitched my own. I figured I could put probably together a Smash Brothers style game in the timeframe, even if I probably couldn't make it look perfect:

  • I've made a platformer before.
  • I got a single controller's input working for Mallow Drop
  • I made a 2D combat system way back at the first MolyJam (even though it was top-down)

I had the pieces for reference, and I could probably fumble my way through handling a second controller. I think by halfway through the first night, I had something playable, even without combat. By Saturday morning, I had something far enough along to warrant going out and buying a second game controller so people could start playing the game. I'd taken a stab in the dark with how the game pad "might work" and when I plugged it in, that gamble paid off - everything worked first time! If fact, the only time I got really stuck this weekend was on a really simple coding error that I was simply too tired to see at the time.

The biggest breakthrough in Towerface by far is the camera/background system - I had thought it would be possible to get something rigged up with flash's "fullScreenSourceRect" mode. I didn't realise how well it would click together. I struggled in getting the camera to focus on the mid-point between the two players, and to get the math figured out, I resorted to using flash's graphics methods to draw lines on the stage as the game played, and as a side effect I ended up playing round with those drawing methods for quite a while, drawing patterns all over the stage as the characters moved around, drawing a line directly between the two characters, drawing a bounding rectangle around them - eventually I settled on the line "bisecting the stage" look that ended up in the game.

So what was the "small scope" that helped so much? Not worrying about the things that didn't help the game.

  • Character sprites were important, but animating them wasn't (especially at 16x16 pixels - there wasn't a lot it would add).
  • Using big blocks of flat colour for the background and making that idea suit the plain black forground - Arting the foreground platforms is something I've had to skip in other games when I ran out of time, but this time I planned to leave it, and spend that time adding variety to the levels instead.
  • Balance - and making the levels each be evocative of some slightly different aspect of arena fighting, needed to be the focus - Some levels (the first ones I added) are based on existing games (Towerfall, Foiled, and Smash Brothers), where some are designed to show off the features and strategies that emerged through playtesting: Long chases across the horizontal maps, the strategies of wall-jumping and stabbing through one-block walls, and the long, floaty jumps for mid-air combat.

Honestly, most of the second half of the jam was spent playing and tweaking. It was Zach - a jammer sitting next to me - who suggested both moving the faces back and forth between the two characters to indicate the score, and to change the game mode from "first to 5 points" to "first to get 5 points ahead". This resulted in more tense back-and-forth moments through the game, and reduced the time needed to create UI and the graphical clutter that would bring.

The final thing that helped control the scope and trim down the dev time? MUSIC. Meghann O'Neill (@Firkraags) joined the jam this year, doing music for 6 games over the two days, and after a quick chat after the pitches (I think I said something like: "I'm thinking something like Looney Toon's chase music / Monaco's soundtrack"). Meghann had the games two themes to me (one for each character) in about 20 minutes. Having that manic feeling to match the character speed to, and design the levels around (the long chase areas for example) sped things up incredibly. Playing the game with the music from minute one forced everything to either fit or be dropped immediately.