Auto-cashout targets and expected value: why no target beats the edge

In a standard crash game every auto-cashout target has the same expected value: you get back about (1 − edge) of what you stake, so with a 1% edge you expect to lose 1% of every unit wagered. The target changes how often you win and how much each win pays, but those two changes cancel out exactly.

This guide shows the cancellation with numbers, then shows what the target really controls: variance.

Expected value in one line

Expected value (EV) is the average result per round if you could play forever. For an auto-cashout at target x with stake 1:

• You win with probability p = (1 − edge) ÷ x and receive x (net profit x − 1).
• You lose with probability 1 − p and lose the stake (net −1).

So the expected payout is:

p × x = ((1 − edge) ÷ x) × x = 1 − edge

The x cancels. Whatever target you choose, the expected payout is 1 − edge per unit staked, and the expected net result is −edge. With a 1% edge that is −0.01 per unit, every time.

The table: same EV, very different rides

All figures per 1 unit staked, 1% edge, computed in Python. "Std. dev." is the standard deviation of one round's result, a measure of how far a single result typically lands from the average.

Read it column by column:

• EV is identical. −0.01 per unit on every row. Over 1,000 one-unit rounds you should expect to be about 10 units down, whatever the target.
• Variance grows with the target. At 1.10x, a 1,000-round session usually lands within about ±10 units of that −10. At 100x the typical swing is ±313 units. That's why high-target sessions occasionally finish far ahead and very often finish far behind.

The swing column is what makes crash feel beatable. On a high target, some players will be well up after a session purely by chance. That isn't evidence that the target "works"; it's the width of the spread.

Worked example: two fictional players, 200 rounds each

The players and numbers below are invented to illustrate the math. Both stake 1 unit per round at 1% edge.

Player Low (target 1.10x). Expected wins: 200 × 0.90 = 180. Each win nets +0.10, so wins add +18.0. Expected losses: 20, each −1, so −20.0. Expected total: −2.0 units.

Player High (target 10x). Expected wins: 200 × 0.099 = 19.8. Each win nets +9, so wins add +178.2. Expected losses: 180.2, each −1, so −180.2. Expected total: −2.0 units.

Same expected result, built in opposite ways. Player Low wins nine rounds in ten but each loss wipes out ten wins. Player High loses nine in ten but each win covers about nine losses. Both end up paying the same 1% on 200 units of wagering.

Why "low and steady" isn't safe

Low targets are often described as a safe approach. The table explains the feeling: the swing is small and wins are frequent. But two things are still true:

1. The loss rate is the same. You pay 1% of everything you wager. Low targets just make it arrive smoothly instead of in lumps.
2. You wager more for the same fun. Because each round moves your balance so little, people tend to play many more rounds at low targets. Expected loss scales with total amount wagered, so more rounds means more expected loss.

Why "high and patient" isn't value

High targets look like the opposite trade: rare wins with big payouts. The EV is unchanged, but the risk of a deep drawdown is much larger. At 10x you will regularly see 20, 30 or more losses in a row (the chance of at least 20 straight losses at 10x is 0.901^20 ≈ 12.4%). And if the site caps payouts per round, high targets become worse than −1%, because the biggest wins are trimmed.

Checklist: choosing a target honestly

If you play, the target is a choice about variance, not value. Before you set one:

• Accept the cost. Expected loss = edge × total wagered. At 1% edge, 500 units wagered costs about 5 units on average.
• Pick the ride you can tolerate. Low targets: frequent small wins, rare full losses. High targets: long losing runs, rare big wins.
• Check payout caps. A cap makes high targets worse than the table.
• Set a session limit in units wagered or time, not in "until I'm up". Chasing a target result means wagering more, which raises expected loss.
• Don't change targets to "recover". Switching after a loss doesn't change EV; it only changes the variance of the hole you're in.

The bottom line

There is no auto-cashout target that turns crash into a positive-expectation game. The formula is built so that payout and probability cancel, leaving the edge on every round. Understanding that is the most useful thing you can know before playing: it tells you the price, and it tells you that no clever target will lower it. The next question people ask is whether a betting system can do what a target can't. It can't either; see martingale and other betting systems on crash. For the odds behind each target, see crash odds math.