RUSH was free speed. We had to make it cost something.

NiceBounce climbs by landing PERFECTs. Chain enough of them and the NICE meter fills; spend it on RUSH and the ball moves faster, the sweet spots shrink, and every floor you touch starts to crumble behind you. On paper that's a classic risk/reward lever. The first version wasn't actually a risk.

A boost nobody should refuse

Early RUSH halved combo points, shrank the sweet spot to 60% and let floors crumble in 0.22 seconds — fast enough that a single mistimed landing during RUSH usually ended the run. Measured against never using it at all, it was a flat downgrade across the climb: strong runs lost score, and mid-skill runs lost nearly half of theirs. A reward nobody should take isn't a choice, it's a trap with a shiny label.

Widening the reward helped the strong player. It didn't fix the trap.

The obvious next move was to make the payoff bigger: faster bounce, double combo points, a more forgiving sweet spot. That helped a skilled player pull ahead, but a mid-skill run was still dying mid-RUSH more often than not — 29 of 55 recorded deaths at that skill level happened during an active RUSH. The reward had changed. The trap hadn't.

The lever was never the reward. It was the crumble delay.

Stretching how long a floor survives after you land on it during RUSH — from 0.22 seconds to 0.9 — is what actually flipped the outcome. Not more points, not a bigger sweet spot: more time to recover from one mistake before the floor under you disappears. With that one number changed, RUSH stopped being a trap with good marketing and became an actual choice.

That's the shape it shipped in: RUSH climbs 10 to 20% faster at every skill level a player can reach, and a miss while it's active is still the single most common way a run ends. Nothing stops anyone from spending the meter the moment it fills. The numbers just make sure that's a decision, not a reflex.