Run martingale, reverse martingale, d’Alembert, Fibonacci or flat betting over thousands of seeded sessions of Dice or Crash. Ruin odds, drawdowns, charts, and why the 1% never moves. Everything runs in your browser on a seeded generator, so a run can be shared and reproduced exactly.
Sponsored
PlinkoLab is sponsored by Craze. Practise here for free until the odds make sense, then decide for yourself whether to play for real.
From Craze's published terms, checked 2026-09-27. Full terms at Craze.
18+. Gambling involves risk. Terms and verification apply. Only where Craze may lawfully serve you.
Pick a strategy, press Run.
Thousands of sessions later you will see how often the bankroll goes to zero, how often it ends above where it started, and how deep the drawdowns go. The seed makes every run reproducible.
Every round here returns 99% of the stake on average: Dice pays 99/p on a p% chance, Crash pays x on a 0.99/x chance, so each round's expected value is −1% × stake. A staking plan only decides how much you stake next; it cannot see the next result. Add the rounds up and the expectation of the whole session is −1% × total turnover, whatever order or size the bets came in. The simulator prints that realised figure for every run; over many sessions it hovers around −1%, and where it does not, the culprit is variance from rare, very large bets, not an edge.
Martingale does not shrink that number; it reshapes the distribution. Doubling after each loss converts a stream of small, likely losses into many small wins and one rare, enormous loss: the streak that outruns the bankroll or the table limit. That is why its median session looks healthy while its probability of ruin is high and its worst drawdown is 100%. Reverse martingale does the opposite trade (many small losses, rare big wins); d'Alembert and Fibonacci sit in between. None of them buys expected value, because there is none to buy.
Play money only. This is a teaching tool, not a staking plan; see how the fairness works.