Crosswalk simulator with the real lane-by-lane multiplier and death probability at each difficulty level.
Crossing games multiply your stake for each lane survived, with the multiplier rising as the survival probability per lane falls. Difficulty selects the hazard rate: easy modes let you cross many lanes for small steps, hard modes pay steeply for two or three.
The simulator shows the cumulative survival probability alongside the multiplier, which is the number the game itself never displays.
Survival probability compounds downward. Ten lanes at 90% each is a 35% chance of getting through, even though every individual step looks safe. That gap between per-step and cumulative odds is what the format is built on.
Deciding your exit lane before you start is the only strategy that survives contact with the game, because the decision gets emotionally harder exactly as the odds get worse.
Each lane is an independent trial, so surviving ten lanes means multiplying ten survival probabilities together. Even at a gentle 95% per lane that is a 60% chance of reaching the tenth — and difficulty settings push the per-lane figure much lower.
The simulator makes the compounding visible, which is the point. A per-step risk that feels trivial produces a cumulative risk that is not, and that gap is where every "just one more lane" decision goes wrong.
It is also the clearest demonstration that a cash-out rule beats an instinct. Deciding the lane you will stop at before the round starts costs nothing here, and it is the only part of the game a player actually controls.
On the implementations we modelled the per-lane probability is fixed by difficulty; what changes is your cumulative exposure.
At operators with a full implementation, yes — the lane outcomes derive from the seed pair and can be recomputed afterwards.