Crash Game Strategy

The swing is the standard deviation of one round's result, measured in stakes. It grows with the square root of the target while the return stays at 97% on every row, which is why picking a target amounts to picking a variance and nothing more.

A very low target wins almost every round and pays almost nothing, so the small, steady charge of the game is never masked by luck. At 1.01x, a player is ahead after 10 rounds 66.8% of the time, after 100 rounds only 1.8%, and after 1,000 rounds under 0.1%.

A high target leaves the result to a handful of wins, and that handful can sit on either side of its average for a long time. At 10.00x, the share of players ahead is 37.7% after 100 rounds and still 35.0% after 1,000. Both players pay the same on average; the difference is whether the cost arrives as a slow, near-certain drain or as long dry spells interrupted by large payouts.

The key number: Over 100 rounds at 97%, 1.8% of players aiming at 1.01x finish ahead, against 37.7% of those aiming at 10.00x, for an identical average loss.

Ahead means a balance strictly above the starting one, with one stake per round. Down a column, targets are compared over the same session. Along a row, low targets fade quickly as the cost builds up, while the highest targets start low and climb, because most short sessions at them contain no win at all.

A losing run counts consecutive rounds that bust before the target. These are the stretches a bankroll has to outlast at a fixed stake, and at the highest targets they reach into the hundreds of rounds, longer than many sessions last.

Many crash games show two bet panels, staked and cashed out separately. Each bet returns the RTP on its own, so the pair returns it too; splitting the stake changes the shape of a round's result, never its price.

Take the familiar pairing of equal stakes at 1.5x and 10x. The round finishes with less money than it started in 90.3% of rounds, because the lower bet on its own returns 1.5 stakes out of the two staked, a net loss of 0.50 stakes. It shows a profit only when the multiplier passes 10x, in 9.7% of rounds, and its swing per stake is 1.60, between the 0.72 of the low bet played alone and the 2.96 of the high one.

With equal stakes, the lower bet covers both only when its target is at least 2x. At exactly that point a round in which only the first bet is paid breaks even, with a net of 0.00, so the round loses money only when it busts before the first target. Below it, every round that stops between the two targets is still a loss, however early the first bet was paid.

Every pair returns 97.000% of the total staked. The pairs whose lower target is 2x never lose a round in which the first bet is paid; all the others lose in every round that ends between their two targets.

An automatic cash-out does not improve the odds of the target it is set to; what it changes is whether that target survives the session. Every figure on this page assumes the same target in every round. A player who decides mid-flight is playing a different target each time, chosen in the heat of the moment, and the swing and losing-run tables stop describing the session actually played. How the controls work, manual clicks included, is explained in the how to play crash guide.

Some games let a player take part of the stake off the table during the flight and let the rest ride. Cashing half at one multiplier and the other half at a higher one is the two-bet case played from a single panel: the same return, with a swing between those of the two targets. It can smooth a session, but it cannot make one any cheaper.

In crash the cost of a session follows the amount staked, and the amount staked follows the pace of the rounds. With a round every 10 seconds a player can place 360 bets an hour, and at 97% a single bet of one stake per round costs 10.8 stakes an hour on average; a second panel doubles that to 21.6. On a version set at 95%, the same pace costs 18.0 stakes for one bet.

Round lengths differ by game and by format. A shared round includes a betting window before each flight, while a solo round starts the moment a bet goes in and can repeat as fast as the player clicks. The table therefore gives the cost at several paces rather than claiming any game's actual timing.

Losses are in stakes at 97%, with one stake per bet and a bet in every round; the paces are examples, not measured timings of any game. A slower pace, a smaller stake or sitting out rounds lowers the hourly cost, and no choice of target does.

Over a long session the average result barely depends on the target. With a bankroll of 100 stakes and 1,000 rounds, players who last to the end lose about 30.0 stakes on average whichever target they hold. What the target decides is how players scatter around that figure.

At 1.2x almost nobody goes broke and almost nobody wins: under 0.1% lose the bankroll, 2.0% finish up, and the middle player ends with 69.6 stakes. At 100x the field splits: 77.6% lose everything and 19.4% finish ahead, every one of them having at least doubled the bankroll, while the median balance is 0.0 stakes.

A shorter session narrows the range. Over 100 rounds at 10x, 37.7% finish ahead and the median balance is still 100.0 stakes, because a short session leaves the cost little time to show. Bet only what you could lose without hardship.

Each session starts with 100 stakes and bets one stake a round at 97%, for 1,000 rounds or until the balance can no longer cover a stake. Higher targets lose more bankrolls outright yet leave more players ahead.