Ball of Steel is now complete as a game. You can open it, change how it controls, play a run, lose it in either of the two ways there are to lose, and go again. Everything after this point is release work rather than design work.
Two decisions did most of the shaping, and both are worth writing down.
Difficulty stacks instead of climbing
The obvious way to make an arcade game harder is to turn a dial: the ball gets faster, the bowl gets steeper, the numbers get bigger. It works, and it is boring, because the thing you are being asked to do never changes — you are just asked to do it more.
Instead, complications arrive one at a time and then stay. Waves start moving across the surface. Debris drops in. The colour drains away. The hole itself begins to widen and drift off centre. The horizon leans.
Each of those on its own is very survivable. The game gets hard because by ninety seconds you are dealing with all of them at once, and because they do not arrive on a fixed rhythm — they come and go on their own schedules, so no two long runs feel the same. The player learns a new problem roughly every twenty seconds instead of learning one problem and then grinding it.
Nothing is allowed to make the bowl unreadable
The counterpart to that decision, and the one that took longer to get right: every complication has a hard limit.
The waves never touch the rim or the edge of the hole. The colour shift never flattens out the sense of depth. The hole never closes on any side, so there is always a way round. The camera never leans past the point where you would lose your bearings.
The rule behind all of them is the same. A player should always be able to see what is happening and still lose. Losing has to feel like a mistake you made, because that is the feeling that makes someone tap play again — and "I could not see it coming" is the feeling that makes them close the app.
What is left before release
Tilt on real hardware. How a tilt control actually feels in two hands is the one question a desk cannot answer. Settings already lets either direction be flipped and the sensitivity changed, so the remaining work is choosing good defaults on real phones.
Music. The sound effects are finished. There is no soundtrack yet, and a game about slow, tense balance needs one.
Store readiness. Listings, screenshots, age ratings and privacy consent for both Android and iOS.
The plan from here is a soft launch on one store first — a small audience, a short list of things to watch, and a real answer on whether one more go is genuinely irresistible — then both stores together.