← all posts site ↗

freeze the keys, free the meanings

the pattern is always the same. i get an image in my head, start building, and the moment the game starts feeling anything close to it, the death spiral begins and the prototype dies. the advice that never worked was about content — think through the cases, plan the mechanics, account for the edge cases. nobody can account for the cases in advance and that was never the job. the job is to stop judging cases and start judging shapes. one rule about decision shape handles any future idea; rules about content keep requiring you to predict the future, which is what kept failing.

the rule that landed is a freeze on keys and modes: the control scheme locks, forever, and a new key or a third mode goes to the sequel file, no exceptions — especially when it feels like part of the game, because that feeling is exactly what the filter exists to kill. but meanings stay free. an existing key can gain a new meaning in a new context whenever it wants. controls can't actually be locked on paper, because verbs get discovered by feel. the lock lands the day the scheme feels right in your hands, and from that day forward a new key costs a sequel file entry. discovering a new key on day two is the prototype working. discovering it on day nine is the disease.

the other half is shipping. the data says it plainly: seventeen prototypes and one release in twelve months. the bottleneck is not making, it's finishing. so the plan is to convert one dead prototype into a public artifact first, with a hard two week cap, before any new game gets six weeks of attention. the order of the builds is the argument.