Crash auto-cashout EV calculator

How Stake Crash works

A multiplier starts at 1.00× and rises continuously. At a random point — drawn from a fair distribution — it crashes. If you cashed out before the crash, you win bet × cashout_multiplier. If you didn't, you lose your bet.

Stake Crash uses a 1% house edge. Mathematically, the probability of the crash exceeding any target multiplier X > 1 is:

P(crash ≥ X) = (1 − h) / X    where h = 0.01

So P(crash ≥ 2.00×) = 0.99 / 2.00 = 49.5%. P(crash ≥ 10×) = 9.9%. P(crash ≥ 100×) = 0.99%.

Why no auto-cashout target beats the edge

For any auto-cashout target T:

EV per bet = P(reach T) × T − 1
           = ((1 − h) / T) × T − 1
           = (1 − h) − 1
           = −h
           = −1%

The expected value per dollar wagered is constant: −1%. Whether you cash out at 1.5× or 50×, the long-run loss rate is identical. What changes is variance.

Variance and longest losing streak

Lower auto-cashout = more wins but smaller. Higher auto-cashout = fewer wins but bigger. The geometric distribution of losing streaks at any target follows:

P(losing streak of L) = (1 − P_win)^L
P(longest streak ≥ L over N rounds) ≈ 1 − (1 − P_lose^L)^N

For 2× auto-cashout (50.5% loss probability) over 100 rounds, the 95% confidence longest losing streak is about 9. For 10× auto-cashout (90.1% loss probability), it's about 30. For 100× (99.01% loss probability), it's about 470.

Translating: at 100× auto-cashout you'll see 470-round losing streaks 5% of the time. Each of those wipes 470 × bet from your bankroll before you hit a single 100× win that recovers it.

"Strategies" that don't work

Common Crash strategies that fail:

• Pattern reading. Past rounds don't predict future rounds. Each round is independent.
• Trailing the average. If average crash multiplier is 2×, no it doesn't mean you should always cash at 2×. The distribution is heavy-tailed; mean is misleading.
• Doubling after losses. Geometric losing streaks at low cashout targets are still long enough to bust most realistic bankrolls.
• Pulling out at high multipliers when "due". A 50× crash is not "due" after 99 sub-50 rounds. The 100th round has the same 1.98% chance as the first.

Practical play

Two reasonable approaches:

1. Low cashout grind (1.2×–2×). Frequent small wins, low variance. Long bankroll life. Good for long sessions where the entertainment is the experience.
2. Moonshot (10×+). Rare hits, big variance. Best with small bet sizes and the understanding that 90%+ of sessions will end at zero. Treat each bet as a lottery ticket.

Both have the same EV. The choice is about what you want from the session, not what gives you better expected returns.