Crash at Shuffle

What Shuffle publishes about Crash

Shuffle's Crash is a multiplayer Original: one curve climbs from 1.00x for the whole table and stops at a point fixed beforehand. Its game page states the price plainly, "Crash runs at a 99% RTP, which is a 1% house edge, and that figure holds whatever cash out target you choose", and prints a specification table (Shuffle Crash game page, 12 Nov 2026).

The $250,000 limit sets the real ceiling

The page is direct about how the two maximums interact: "a curve that runs into six figure multiples will still pay out at the per round cap, so the ceiling on your stake is the payout limit rather than the multiplier". The help centre adds the mechanism, and it is specific to this game. For most Shuffle Originals a bet that could exceed the maximum win is refused. "Crash and Stonks work differently: your bet will still be placed, but if your winnings reach the maximum win, your bet will be automatically cashed out at the maximum win amount."

So every Crash bet carries an automatic cash-out at $250,000 divided by the stake, whether or not you set one. The table shows where that falls and, taking the stated 99% return at every target, how often a round gets there.

Arithmetic from the $250,000 maximum win and $10,000 maximum bet (Shuffle help centre, 12 Nov 2026) and the stated 99% return, under which a round reaches a multiplier t with probability 0.99 / t.

At the maximum bet the automatic exit is at 25x, a level reached in about 4% of rounds. The 1,000,000x on the specification table is collectable in full only by a stake of 25 cents or less. For anyone betting more than a few dollars, the multiplier that matters is the one in the first two columns.

How often a target is reached

If the return is 99% at every cash-out target, as Shuffle states, the chance that a round reaches a target t must be 0.99 divided by t. Nothing else gives a flat 99%.

Arithmetic from the stated 99% RTP at every target.

A 2x target fails slightly more often than it succeeds: 50.5% of rounds stop short. A 1.5x target wins about two rounds in three and loses the full stake in the third. Shuffle's page describes the same trade without the numbers: "A 1.5x cash out hits often and returns little; a 50x cash out will not arrive for a long time and pays accordingly. The edge is the same at both". Its Trends chart of recent rounds is, in the page's words, "a record of what happened and not a signal about what is coming".

Shuffle allows several bets on one round, and the page suggests running "a low target and a high target on the same round". Each of those bets returns 99% on its own, so the pair returns 99% of the combined stake. Splitting changes the shape of results and leaves the cost where it was.

Two descriptions of the fairness model

Shuffle describes Crash fairness in two places, and they do not say the same thing. The game page uses the language of its single-player games: "Every round's crash point is generated from a server seed, your client seed, and a nonce", with a client seed you "can rotate at any time". The help centre's "Provably Fair - Game Events" article describes something else for Crash: "There are 2 parts to the calculation of the result of a crash game", a Bitcoin block hash and a seed from a pre-generated chain.

The help-centre version is the one that fits a game where every player sees the same curve, and it is the one with detail that can be checked:

• The chain. "The seed for the 10 millionth game is randomly chosen by Shuffle", and "the seed for each game is the result of hashing the seed of the subsequent game". Seeds are therefore used in the reverse of the order they were generated, and each revealed seed hashes to the one before it.
• The commitment. "By publishing the hashed value of the hash used in the first game, players can independently verify the integrity of the hash for all games up to the 10 millionth game."
• The salt. The hash of Bitcoin block 779588, 000000000000000000033ab71b5fa7edcafe388e5f49719f1de8e1f8d0670460, announced in a seeding post on X that the article links. That block has a time of 13:30 UTC on 6 March 2023.
• A worked link. The article gives game 1323185 with hash 2f824ebb…825ac0 and game 1323184 with hash 20f42ccc…280265, and states that the second is the SHA-256 of the first.

What the article does not give is the last step: the formula that turns a game's seed and the block hash into a multiplier. The chain and the salt are documented; the conversion is not. A player can therefore prove from Shuffle's documentation that a round's seed belongs to the committed chain, and has to rely on Shuffle's own verifier for the crash point itself. Slide, Shuffle's other shared-round game, "uses the same provably fair and seeding process as Crash" with Bitcoin block 952294 as its salt.

Verifying a round

1. Open a finished round from the game's history and copy its hash.
2. Hash that value with SHA-256. The output must equal the hash shown for the round immediately before it. The help centre suggests any online SHA-256 calculator for this.
3. Repeat backwards as far as you like. Every link that holds is a round whose seed was fixed before it was played.
4. For the crash point, use the verification page in Shuffle's Provably Fair section, which takes the round's hash.

Our Shuffle seed checker and the Crash verifier do steps 2 and 3 for Shuffle: paste a game hash and they print the SHA-256 that the previous round's hash must match. They do not print a Shuffle crash point, because the conversion is not in Shuffle's documentation. For Stake, whose formula is published, the same tools compute the point as well.

The proof has a boundary. It shows that crash points were set in advance and cannot react to the bets in a round. It does not cover the path between your click and Shuffle's server, where a manual cash-out sent a moment too late is a loss. An auto cash-out is decided before the round starts and cannot arrive late, which is the practical argument for the control Shuffle's page calls "the most useful control in the game and the one most players ignore". Background: what provably fair gambling proves.

Shuffle Crash against Stake, Roobet and Gamdom

Stake: BitcoinTalk seeding post (read 12 Nov 2026), help centre 12 Nov 2026. Roobet and Gamdom: verifier code linked from each fairness page, 12 Nov 2026; Gamdom limits from its game page, 1 Nov 2026.

Shuffle states the same 99% as Stake and documents its chain and salt in its help centre. The gap is the unpublished conversion step, which leaves the 99% as a stated figure that cannot be re-derived from the documentation, where Stake's can. Shuffle's limits are lower than Stake's or Gamdom's, and it is the clearest of the four about what happens when the limit is reached.

Sources: Shuffle Crash game page (12 Nov 2026); Shuffle help centre, maximum bet and win limits and game events (12 Nov 2026); Bitcoin block 779588. Probabilities and limit arithmetic are our calculations from the stated 99% return and the published limits.

The auto cash-out calculator gives hit rate and expected loss for any target. Shuffle's licence, KYC and withdrawal terms are in the Shuffle review.

Frequently asked questions

Related pages