Ticket by ticket
What happened in our three-week test. Each feature built to "done", one after another.
The same four people and the same three weeks, building Space Invaders for a publisher showcase on Friday 25 September. On the left, what happened in our test, one ticket at a time. On the right, the same work done bronze first: the whole game playable early, then upgraded.
What happened in our three-week test. Each feature built to "done", one after another.
The same people and effort, planned the other way. Everything in at bronze, then upgraded where players look.
Every feature, every working day. Ticket by ticket, the game is never playable from start to finish within the three weeks. Bronze first, the whole game is playable by the fourth day, and the rest of the time goes into making it better.
No To do, In progress and Done. Each part of the game sits in one of four states, checked every night by playing the build against the design. Medals are the studio's words; the outside world hears the plain ones.
Not in the build.
Works end to end with placeholder art, sound and feel.
Does what the design says, at the quality it describes.
What you'd put in the trailer.
Bronze isn't failure, it's step one. Being behind means more bronze, never crunch.
The question stops being "will we finish?" and becomes "how good will it be?", answered every night from the build itself.
You always get a playable build on the date. What we report is how much of it is at final quality.
A missed milestone stops being a risk you discover late. It becomes a quality level you choose early.
Where the numbers come from. The left side is the record from our three-week test: an invented four-person team (code, art, design and a remote sound designer) whose real start and finish dates for 26 tasks were mapped onto 15 features of the game. On the showcase day, win and lose conditions had never been started and high-score saving had been cut, so the game could be played but not won or lost.
The right side is an illustrative plan, not a measured result. It assumes the same total effort, with a bronze version costing about a third of a finished one and gold about half as much again as silver. Spending effort on bronze everywhere first means fewer features reach "as designed" (10 at silver or better, against 12 finished on the left), in exchange for a complete, playable game from day 4 and nothing missing on the showcase day.