Shipping small is a training plan

Every team I've worked on has had the same temptation: pause, rewrite, come back stronger. It almost never pays off, and the reason is the same reason runners blow up in week two of a new plan.

Progress compounds from repeatable units of work. A pull request you can review in ten minutes is an easy run. A three-week refactor is a max-effort session you didn't earn the base for.

The practical rule I use: if a change can't be described in one sentence and reverted in one command, it's too big. Split it until both are true.

This isn't a productivity trick. It's load management. Consistency beats intensity over a quarter, the same way it does over a training block.