Knowing when to stop iterating

Games are never finished, only released. Every published designer will tell you they could have kept tweaking forever, and that at some point they chose to stop. The practical question is not "is the game perfect?" but "have the changes stopped mattering?" When successive versions differ only in wording tweaks and one-point adjustments, and playtest feedback has shifted from problems to preferences, the design work is essentially done.

There are recognizable signs of convergence. Blind playtest groups get through full games without rules questions. New testers' complaints stop clustering: instead of five people hitting the same wall, you get five different mild opinions, which is taste rather than defect. Your changelog between versions gets shorter. Testers start asking to play again without being invited, and a few ask where they can get a copy. When the same version survives eight or ten plays across different groups without generating a change you feel compelled to make, you have arrived.

Beware the two failure modes on either side. Stopping too early usually looks like avoiding blind tests and strangers, so the game never faces the conditions that would expose remaining problems. Iterating too long usually looks like redesigning in response to individual taste, chasing the approval of every single tester, or repeatedly tearing up systems that work because a new idea seems more exciting. That last pattern sometimes means the new idea deserves to be its own game, not a renovation of this one.

Done also depends on destination. A game headed to publishers needs a stable, blind-tested design and a clear pitch, but publishers expect to do further development themselves, so pixel-perfect balance is not required. A self-published game needs to be genuinely finished, because no developer stands between your files and the customer. Either way, set a concrete exit condition, such as three consecutive blind tests with no rules failures and no repeated complaint, and honor it when you get there. Shipping is a skill, and it is only learned by shipping.