Crash Game Math: From a Hash Chain to Expected Return

A rising multiplier makes Crash look like a timing problem. The mathematics asks a different question: how often does a round reach a chosen target, and what does that probability imply for the return on a stake?

In Crash, the multiplier starts at 1.00× and rises until the round ends. Cashing out before the crash pays stake multiplied by the cashout multiplier; remaining in the round when it crashes loses the stake. Players in the same round share its result.

The model below is the published bustabit model described in the original CryptoPlay Guide article. Its 1% house edge is an assumption tied to that formula. Other operators may use different edges or rules, so this is not a universal specification for every game called Crash.

Begin with the target probability

For this model, the probability of reaching a target multiplier is:

P(result ≥ m) = 0.99 / m

The practical consequence is that a higher target pays more when reached but is reached less often. For example:

Cashout target	Chance the round reaches it	Expected return per 1 USDT
1.01×	98.02%	0.99
1.5×	66.00%	0.99
2×	49.50%	0.99
3×	33.00%	0.99
5×	19.80%	0.99
10×	9.90%	0.99
100×	0.99%	0.99

The constant expected return is the important relationship. Raising the cashout target changes win frequency and volatility, while the expected return remains the same under the stated model.

For a fixed target, the expected net result per unit staked is:

m × 0.99/m − 1 = −1%

Auto-cashout can select the target. It does not remove the factor responsible for that negative expectation.

Trace the result back to the hash chain

Crash games usually use a hash chain rather than separate per-player seed pairs. In this model, each game hash is the SHA-256 hash of the next one. The operator publishes the final hash and a salt in advance, then plays through the chain backwards.

That gives verification something concrete to check: the relationship between a revealed game hash and the previously committed chain value.

The bustabit calculation is:

h = HMAC_SHA256(key = salt, message = game hash)
X = (first 52 bits of h) / 2⁵²
(uniform in [0, 1))
result = max(1, floor(99 / (1 − X)) / 100)

Notice both the key placement and the mapping after the HMAC. The salt is the key, and the game hash is the message. The calculation takes the first 52 bits, then applies the multiplier formula and rounding.

The 99 in that formula, rather than 100, supplies the model's 1% house edge. A reproducible result can therefore be entirely consistent with a negative expected return for the player.

Check the commitment and the multiplier separately

Verification has two parts. First, hash the game hash with SHA-256 and compare it with the previous round's hash. A matching chain relationship checks that the round fits the committed sequence.

Second, recompute the multiplier using the game hash, public salt and documented formula. This checks the transformation from the committed input into the displayed result.

CryptoPlay Guide's browser verifier includes a Crash mode at https://cryptoplay.guide/tools/provably-fair-verifier/. Its presence does not remove the need to check which model the operator uses. Reproducing the wrong model is not verification of a different operator's rules.

A matching chain and multiplier show that a result follows the commitment and calculation. They do not predict the next crash or establish that an operator will reliably pay withdrawals.

Rounding matters near the start

Under this formula, about 1.98% of rounds end at 1.00×. This follows from the condition 99/(1−X) < 101: results below 1.01× round down to 1.00×.

That behavior belongs to the specific formula, including its rounding. When comparing a mathematical model with a live game, read the fairness documentation and the cashout rules together. CryptoPlay Guide's safety guide also calls out auto-cashout, ties and maximum-win rules as checks to make before relying on a formula.

A general probability calculation cannot establish those operating details by itself.

Changing stakes does not change the edge

A loss-recovery betting system changes how much is exposed on later bets. It does not change the expected return of the underlying game.

Consider Martingale at 2×: eight consecutive losses, approximately 0.42% from a given start in this model, multiply the stake by 256×. The increased exposure is part of the system, while the bets still carry negative expectation.

Likewise, a simulation or losing-streak calculator is a way to inspect a model. It cannot identify the next round's crash point. Choosing a higher multiplier or a different staking sequence does not turn verification into forecasting.

The useful conclusion is modest: inspect the commitment, reproduce the formula, and evaluate the house edge separately. Set an affordable loss limit and session time before playing. A cashout target controls a round; a spending limit controls how much you have decided to risk.

Originally published at https://cryptoplay.guide/guides/crash-game-math-explained/

CryptoPlay Guide, the site this was originally published on, earns affiliate commissions from some casinos it covers. This post contains no affiliate links.

18+ (or the higher legal age where you live). Gambling involves risk of loss; only play where legal, set deposit, loss and time limits, and seek help if gambling causes harm, for example from BeGambleAware.org.