- You win if you cash out before the multiplier crashes. The crash point is fixed before the round starts.
- With a 1% house edge, the chance of reaching a multiplier M is 0.99 ÷ M: 49.5% for 2x, 9.9% for 10x, 0.99% for 100x.
- Every cash-out target has the same long-run return. Your target only changes how bumpy the ride is.
- Martingale does not beat the edge. In our worked example it wins $1 most of the time, but loses $1,023 about once every 927 cycles and still costs 1% of everything wagered.
- Good crash games publish a hash chain so you can verify every round after it ends.
How a crash round works
- During a short betting window, players place bets for the next round. Most games let you set an automatic cash-out multiplier in advance.
- The round starts at 1.00x and the multiplier rises, slowly at first and then faster.
- At any moment you can cash out, locking in your stake times the current multiplier.
- At a random point the multiplier crashes. Anyone who has not cashed out loses their stake for that round.
The crash point is decided before the round begins, not while the line is climbing. The animation is a reveal of a number that already exists, which is what makes it possible to verify afterwards.
Crash is usually multiplayer. You see other players’ bets and cash-outs in real time, which is a big part of its appeal and the reason it is popular on streams. It does not change the maths: every player faces the same crash point and the same edge.
Where the crash point comes from
Most crypto crash games turn a random number between 0 and 1 into a crash point with a formula that builds in the house edge. The open-source crash site bustabit publishes a typical version: it takes a uniform random value X between 0 and 1, calculates 99 ÷ (1 − X), rounds down, and divides by 100, with a minimum of 1.00x. The 99 in that formula is the 1% edge.
That gives a neat result. The probability that a round reaches at least M is exactly 0.99 ÷ M for any target from 1.01x upwards. At 2x that is 0.495, or 49.5%. At 100x it is 0.0099, or 0.99%.
Under this formula, a round reaches 1.01x with probability 0.99 ÷ 1.01 = 98.02%. The rest, about 1.98% of rounds, crash at 1.00x and every bet in them loses. That is not a separate trick added on top: it is simply the bottom of the same curve. It is why even a 1.01x target is never a sure thing.
Other casinos use slightly different formulas, edges or caps on the maximum multiplier, so always read the game’s fairness page. The shape of the curve, roughly “edge-adjusted 1 ÷ M”, is common to almost all of them.
The odds of reaching each multiplier
| Cash-out target | Chance the round reaches it | Rounds per hit, on average |
|---|---|---|
| 1.01x | 98.02% | About 1.02 |
| 1.10x | 90.00% | About 1.11 |
| 1.50x | 66.00% | About 1.52 |
| 2x | 49.50% | About 2.02 |
| 3x | 33.00% | About 3.03 |
| 5x | 19.80% | About 5.05 |
| 10x | 9.90% | About 10.1 |
| 20x | 4.95% | About 20.2 |
| 50x | 1.98% | About 50.5 |
| 100x | 0.99% | About 101 |
| 1,000x | 0.099% | About 1,010 |
| 10,000x | 0.0099% | About 10,101 |
Multiply any row’s target by its probability and you get the same answer: 0.99. Whether you aim for 1.5x or 100x, you get back 99 cents of every dollar wagered over the long run. What changes is how often you win and how large each win is.
What the multiplier distribution looks like
The curve is heavily skewed. Half of all rounds crash below about 1.98x, because 0.99 ÷ 1.98 = 0.5. Roughly one round in ten reaches 10x and roughly one in a hundred reaches 100x. Very large multipliers do happen, but they are rare enough that a short session can easily pass without one.
- Median crash point: about 1.98x.
- Rounds reaching 10x: about 9.9 in every 100, on average.
- Rounds reaching 100x: about 1 in every 101, on average.
- Ten rounds in a row below 2x: 0.505¹⁰ ≈ 0.108%, or about 1 in 927 for any given run of ten. Over 1,000 rounds, the chance of seeing at least one such streak somewhere is roughly 41%; over 2,000 rounds it is about 66%.
That last point matters. Streaks that feel impossible in the moment are routine over a long evening of fast rounds. The history bar showing a string of low crashes is exactly what the maths predicts, and it tells you nothing about the next round.
House edge and RTP across crash games
Crash games at crypto casinos commonly advertise a house edge of around 1%, which is a return to player (RTP) of 99%. It is not universal. Third-party crash titles are usually set lower: Pragmatic Play lists Spaceman at 95% RTP on its own game page, and Spribe’s Aviator is widely listed at 97%. Some studios offer several RTP versions, and the casino picks which one to run.
| RTP | House edge | Expected loss per $1,000 wagered | Chance of reaching 2x |
|---|---|---|---|
| 99% | 1% | $10 | 49.5% |
| 97% | 3% | $30 | 48.5% |
| 95% | 5% | $50 | 47.5% |
Wagered is the key word. Crash rounds are short, so a $10 stake over 200 rounds is $2,000 wagered. At 99% RTP the expected cost is $20; at 95% it is $100. The difference between games matters far more than any cash-out choice. Check the RTP on the game’s info screen before you play, and compare it between casinos. Our RTP guide explains how small edge differences compound.
Auto cash-out and auto-bet
Auto cash-out lets you set a target before the round starts, for example 2.00x. If the multiplier reaches it, the game cashes you out at exactly that value. If it crashes first, you lose the stake. Most games also let you cash out manually before the target.
- It removes reaction time. Manual cash-outs depend on your click and your connection; auto cash-out is processed by the server at the target.
- It does not change the odds. A 2x auto cash-out wins 49.5% of the time at a 1% edge, exactly like a perfectly timed manual click at 2x.
- Partial cash-out. Some games let you take part of the stake out early. Pragmatic’s Spaceman, for example, offers a 50% auto cash-out option. Each part is simply a separate bet at its own target, with the same edge.
- Auto-bet. Many games can repeat bets automatically, sometimes with rules such as “double after a loss”. This is where Martingale usually hides. Set a stop-loss and a number of rounds if you use it, and never let it run unattended.
Auto cash-out is the best responsible-gambling feature crash has. Deciding your target before the round, rather than while the line is climbing and the chat is cheering, removes most of the impulsive decisions that drain balances.
Low and high cash-out targets
A low target such as 1.2x wins most rounds and loses rarely, but each loss wipes out several wins. A high target such as 20x loses most rounds and wins rarely, but each win is large. Both have the same expected return.
Player A bets $1 a round at 1.2x for 100 rounds. Each round wins $0.20 with probability 82.5% or loses $1 with probability 17.5%. Player B bets $1 at 20x for 100 rounds, winning $19 with probability 4.95% or losing $1 with probability 95.05%. Both have an expected result of −$0.01 per round, or −$1 over 100 rounds. But the typical swing (one standard deviation) over 100 rounds is about $4.56 for Player A and about $43.38 for Player B. And Player B has about a 0.6% chance of going all 100 rounds without a single win.
In practice, your target is a choice about volatility, not profit. Low targets give longer sessions with smaller swings; high targets give short, streaky sessions where your balance can fall quickly before a win arrives. Choose a target that fits your budget, not the one that looks best on a stream highlight.
Session maths: how long a budget lasts
Because crash rounds are short, the number of rounds you play matters as much as the stake. The edge works on the total amount wagered, while luck works on the swings around it. Over a single evening, the swings are usually much bigger than the edge, which is why sessions feel unpredictable even though the long run is not.
Suppose you play 300 rounds in a session at $1 a round. That is $300 wagered. At 99% RTP the expected cost is 1% of $300 = $3; at 95% RTP it is $15. Now the swings: at a 2x target one round has a standard deviation of about $1, so 300 rounds have about √300 × $1 ≈ $17.32. At 1.2x it is about $7.90, and at 20x about $75.14. So at 2x, finishing $17 up or $17 down is ordinary, while the expected cost is only $3. Over many sessions the swings average out and the cost does not.
- Budget in rounds, not just money. Decide how many rounds you will play as well as how much you will spend.
- Smaller stakes stretch a session. Halving your stake halves both the expected cost and the swings.
- Higher RTP matters most for long sessions. The difference between 99% and 95% is small per round but large over thousands of rounds.
- Check the maximum win. Many crash games cap the maximum multiplier or the maximum payout per bet. Read the rules so a huge multiplier does not pay less than you expect.
Why Martingale fails: a worked example
Martingale means doubling your stake after every loss, so that the first win recovers all previous losses plus one base stake. At a 2x target it feels safe, because you usually recover. The numbers below show why it is not.
| Losses in a row | Next stake | Total staked so far | Chance of a losing run this long |
|---|---|---|---|
| 1 | $2 | $1 | 50.5% |
| 3 | $8 | $7 | 12.9% |
| 5 | $32 | $31 | 3.28% |
| 7 | $128 | $127 | 0.84% |
| 8 | $256 | $255 | 0.42% |
| 10 | $1,024 | $1,023 | 0.108% |
You start with $1,023, enough for ten doubling bets: $1 + $2 + $4 + … + $512 = $1,023. A cycle ends when you win (profit $1) or lose ten in a row (loss $1,023, and you cannot afford the $1,024 eleventh bet). The chance of ten losses in a row is 0.505¹⁰ ≈ 0.108%, about 1 in 927. Expected result per cycle: 0.99892 × $1 − 0.00108 × $1,023 ≈ $0.999 − $1.104 = −$0.105. The average amount wagered per cycle is about $10.46, so the loss is 1% of money wagered, exactly the house edge. Over 100 cycles the chance of hitting the losing streak at least once is about 10%; over 1,000 cycles, about 66%.
The pattern is the same at any size. Martingale converts many small wins into a rare, very large loss. It cannot change the 1% edge, because every individual bet still has it. Table maximums make things worse: if the game caps stakes, you may hit the limit before your bankroll runs out, and the sequence breaks earlier.
Other systems
- Reverse Martingale (doubling after a win) has the opposite profile: frequent small losses and occasional large wins. Same edge.
- Reading the history. Rounds are independent. A run of low crashes does not make a high one “due”.
- Two bets at two targets. Splitting a stake between a low and a high cash-out smooths the ride a little. It does not change the edge on either bet.
- Waiting several rounds before betting changes nothing, because the rounds you skip have no effect on the next one.
Bots, scripts, Telegram channels and “crash predictor” apps that claim to forecast rounds are scams. If a crash game could be predicted, it would not be provably fair. Many of these tools exist only to take a fee or steal wallet access.
How to verify a crash round is fair
Most crypto crash games use a hash chain. The casino generates a long sequence of hashes in advance, where each round’s hash is derived from the next one, and publishes the final hash before the game starts. Because each hash depends on the one after it, the casino cannot alter a future round without breaking the chain.
Many games also mix in a public value that did not exist when the chain was committed, usually a future Bitcoin block hash. bustabit did this in its fourth seeding event in September 2025, naming the hash of Bitcoin block 917,000 as its salt before that block was mined. Neither the casino nor the players could know the outcomes in advance.
- Find the game’s fairness page and note the published formula and any public salt.
- Copy the hash for a round you played from the round history.
- Run it through the formula in an independent verifier or a short script.
- Confirm the result matches the crash point shown.
- Hash the round’s hash with SHA-256 and confirm it gives the previous round’s hash, proving it belongs to the committed chain.
Third-party crash games vary. Some use their own provably fair scheme, such as combining a server seed with client seeds from the first players to bet; others rely on a lab-certified RNG and cannot be verified round by round. Our provably fair guide explains the general system with a worked example.
Popular crash games and originals in 2026
| Game | Made by | Fairness method | RTP to expect |
|---|---|---|---|
| Aviator | Spribe | Provably fair, server seed plus players’ client seeds | Widely listed at 97% |
| JetX | SmartSoft Gaming | Check the provider’s game rules | Varies by casino configuration |
| Spaceman | Pragmatic Play | Lab-certified RNG | 95% on the provider’s page |
| In-house crash | The casino itself | Hash chain, usually with a public salt | Often 99%; check the fairness page |
| Limbo | Usually in-house | Seed-based provably fair | Often 99%; check the fairness page |
In-house crash and limbo are standard at originals-focused crypto casinos. Operators we review with their own originals include Stake, Roobet, BC.Game, Shuffle and newer brands such as Yeet and Duel. Edges and maximum multipliers vary by casino and by game, so compare the fairness pages rather than assuming they are identical. Our Stake alternatives guide compares several of these operators.
Availability also depends on where you live. In Brazil, Provisional Measure 1,394 of 25 September 2026 banned fixed-odds betting and online casino games, including crash games and offshore operators serving Brazilian players, and licensed sites went offline on 6 October 2026. Our Brazil guide tracks the latest position. Wherever you live, check that online casino games are legal for you before playing, and never use a VPN to get around a restriction.
Crash, limbo and other multiplier games
| Game | How it works | Same maths? |
|---|---|---|
| Crash | Shared multiplayer round; cash out before the crash | Yes |
| Limbo | Pick a target; a single-player result is drawn instantly | Yes, without the live round |
| Aviator-style games | Third-party crash games with a plane or rocket theme | Similar curve, but edge and fairness method vary by provider |
| Dice | Pick a win chance and payout; one roll decides | Same edge logic, different presentation |
Limbo is crash without the waiting: you choose a target and the result appears immediately. Because it plays much faster, the same budget lasts far fewer minutes, and the amount you wager per hour, which is what the edge eats into, rises sharply.
Common crash mistakes
- Chasing a big multiplier after watching one on stream. Highlight clips show the 1-in-1,000 rounds, not the 999 before them.
- Raising the target after a loss. It does not make the next round more likely to go high.
- Cashing out manually on a slow connection. If you rely on clicks, latency can cost you; auto cash-out is processed at the target.
- Ignoring the RTP. Playing a 95% crash game instead of a 99% one costs five times as much per dollar wagered.
- Leaving auto-bet running. Scripts with doubling rules can wipe a balance in minutes once a losing streak arrives.
- Paying for signals or predictors. These are scams without exception.
- Playing with funds you need for something else. Crash is entertainment with a price, not a way to make money.
Most of these come down to the same thing: decisions made in the moment rather than before the session. Choosing your casino carefully helps too. Our crypto casino licence guide explains what to check before you deposit, and our instant withdrawal guide covers getting winnings out quickly.
Bonuses, wagering and crash
Many casinos exclude crash and other originals from bonus wagering, or count them at a reduced rate, because low targets let players clear wagering with very little risk. Breaking a bonus rule can void winnings. Read the game-weighting table before you play crash with bonus funds. Our bonus guide explains how weighting works, and rakeback is often a better fit for high-volume crash players because it returns a share of the edge regardless of results.
Playing crash responsibly
Crash is fast, social and built to keep you watching for one more round. That combination makes it easy to lose track of time and money.
- Set a session budget and a deposit limit in your account before you start.
- Use auto cash-out so decisions are made before the round, not in the heat of it.
- Do not increase stakes to recover losses, whether by hand or with auto-bet rules.
- Decide in advance how many rounds you will play, since rounds are short and add up quickly.
- Take a break after a big win as well as after a loss.
If crash or any other game stops feeling like entertainment, our responsible gambling guide lists the tools and support available, and our safer play hub links to free help services. For casinos with strong limit tools, start from our crypto casino rankings.
Sources
18+ only. Gambling involves risk and should be treated as paid entertainment, never as income. Check that online gambling is legal where you live before you play. Get help and set limits →