Crash Game Maths: Why No Auto-Cashout Beats the Edge

How the multiplier is distributed

A provably fair crash game publishes its formula, and the shape is consistent across implementations. A hash is converted to a uniform random number, and that number is mapped to a crash point so that the probability of reaching any multiplier m or higher is roughly proportional to 1/m, with a small adjustment that produces the house edge. In other words: reaching 2× happens about half as often as reaching 1×, reaching 10× about a tenth as often, reaching 100× about a hundredth as often.

That 1/m shape is what makes the game feel the way it does, and it is also what makes strategy futile. Consider the two obvious bets. Cash out at 2× and you win roughly half the time, doubling your stake. Cash out at 10× and you win roughly a tenth of the time, multiplying by ten. The product of probability and payout is the same in both cases, and in both cases it is slightly below your stake. The shortfall is the edge.

Where the edge actually comes from

The adjustment that creates the edge usually takes one of two forms. Some games reserve a share of rounds as an immediate bust at 1.00×, where no cashout is possible at all; if that share is one round in a hundred, the game keeps one per cent of every wager regardless of strategy. Others shade the distribution slightly so that every multiplier is a touch less likely than the clean 1/m curve would give. Either way the result is a constant: a fixed percentage of turnover, applied identically to the player cashing out at 1.2× and the player hunting 50×.

This is worth stating plainly because crash games are frequently marketed as low-edge, and the figure quoted is usually accurate. A one per cent edge is genuinely better than most slots. What it is not is a figure that strategy can move.

What the cashout point does change

If expectation is fixed, the choice is not meaningless — it determines variance, and variance determines what your session actually feels like and how long your bankroll survives.

A low cashout at 1.2× or 1.5× wins most rounds and loses a large multiple of a single win when it loses. The balance graph grinds upward and then drops sharply. Players describe this as reliable, which it is until it is not: a run of six busts below 1.2× is an ordinary occurrence, and it removes six wins' worth of profit.

A high cashout at 10× or above loses almost every round and occasionally returns a large multiple. Long losing sequences are not bad luck here, they are the design — the median outcome of a hundred rounds at 10× is a loss, and the mean is carried by the few hits.

The risk of ruin differs sharply even though expectation does not. With a fixed stake and a finite bankroll, a strategy with higher variance has a higher probability of hitting zero before any given number of rounds, which is the one practical sense in which the choice matters.

Why no auto-cashout at all is not worse

Manual play has the same expectation as any automated point, because every exit you might take is already on the same curve. What it has instead is an extra failure mode: latency, hesitation and the temptation to hold on. If you are going to play, the automated version of your own plan executes it without those, which is an argument for auto-cashout as discipline rather than as edge.

The strategies that circulate, and what each one actually is

• Martingale on crash. Doubling after a loss at a 2× target converts many small wins into one catastrophic loss. Expectation unchanged; variance enormously increased; ruin probability near certain over enough rounds, bounded only by the table maximum, which exists for that reason.
• Two simultaneous bets, one low and one high. A popular pattern and a pure blend: the expectation is the weighted average of two identical expectations. It smooths the graph, nothing more.
• Waiting for a run of low crashes before betting. This is the gambler's fallacy with a hash function underneath it. Each round's seed is independent of the last, which is the whole point of the provably fair construction — and the published verification lets you confirm independence rather than assume it.
• Following the live multiplier history panel. The panel exists because patterns are compelling to look at. The data in it has no predictive content.

What to check in the game you are actually playing

Since the edge is the only variable that matters, read it off the game rather than the marketing.

• Find the published formula or fairness page and see which edge mechanism it uses — a reserved instant bust, a shaded distribution, or a stated house-edge percentage.
• Check whether the stated edge applies to the base game only. Side features, insurance options and jackpot contributions usually carry a worse rate, and their cost is added to yours.
• Check the maximum multiplier and the maximum win. A capped maximum truncates the top of the distribution, which raises the effective edge for anyone targeting high multipliers.
• If a bonus is wagering, check the game's contribution and any maximum-bet rule. Crash games are frequently excluded or reduced, and that changes the arithmetic far more than any cashout point.

Gambling is for adults and a negative expectation is the design rather than a flaw in it; the mathematics above says that no cashout strategy escapes it. If a system has started to feel like a reason to keep playing, the free services on our responsible gambling page are independent of any operator.