Crash game odds: Crash, Limbo, Aviator and Chicken Road
Divide the RTP by the multiplier you're aiming for and you have your chance of getting there, in Crash, Limbo, Aviator or Chicken Road. At 99% RTP, a x2 cash out comes off in 49.5% of rounds. Push it to x10 and you're at 9.9%, about one round in ten.
The formula behind every crash game
A crash round only ever produces one number, the point where the curve stops. Your bet wins if that point gets to your target. The 99% has to hold at every target, so a x5 target, which pays five times the stake, can only come in 19.8% of the time. x3 works out at 33%, x50 at 1.98%.
| Target | Chance at 99% RTP | About 1 in | Chance at 97% RTP |
|---|---|---|---|
| x1.01 | 98.02% | 1.02 | 96.04% |
| x1.10 | 90.00% | 1.11 | 88.18% |
| x1.50 | 66.00% | 1.52 | 64.67% |
| x2 | 49.50% | 2.02 | 48.50% |
| x3 | 33.00% | 3.03 | 32.33% |
| x5 | 19.80% | 5.05 | 19.40% |
| x10 | 9.90% | 10.1 | 9.70% |
| x50 | 1.98% | 50.5 | 1.94% |
| x100 | 0.99% | 101 | 0.97% |
| x1,000 | 0.099% | 1,010 | 0.097% |
| x1,000,000 | 0.000099% | 1,010,101 | 0.000097% |
Own calculation, chance = RTP ÷ target. The 97% column is an example for games that list a lower RTP. Read the actual figure in the rules of the game you play.
Limbo is Crash without the wait
Limbo's formula sits on Stake's provably fair page as plain code. It takes a random float, divides 0.99 by it, cuts the result to two decimals and shows anything under 1 as 1.00. You set the target first and the number appears at once, with no curve and no cash out button.
We put that formula through two million rounds in Python. x2 or better came up 49.564% of the time against 49.5% on paper, and x100 came up 0.994% against 0.99%. There's a floor too. 1.98% of rounds land under x1.01, so even a x1.01 target loses about once every 50 bets.
| Target | Simulated share at or above | Formula |
|---|---|---|
| x1.5 | 66.060% | 66.000% |
| x2 | 49.564% | 49.500% |
| x10 | 9.897% | 9.900% |
| x100 | 0.994% | 0.990% |
2,000,000 rounds of Stake's published Limbo formula, Python random, seed 612. Script: growth6/daten/phase2/b1a/crash.py.
Aviator and Chicken Road
Aviator flies a plane where Stake has a rocket, and the maths underneath doesn't change. Chicken Road feels different because you walk the chicken across one lane at a time and decide after each lane whether to stop. Every stop is still a cash out at a fixed multiplier. If the game pays the same RTP whatever you do, chance times payout comes back to that RTP at every lane.
Get that number from inside the game, on the rules or info screen. Review sites copy each other, and we found different figures for both games without a studio page we could actually load to settle it. One point of RTP moves every chance in the table by about 1% of its value.
The doubling trap
Somebody in chat will always suggest doubling after each loss at x2. Most runs end one unit up, and that keeps going until a streak arrives that the balance can't pay for. Ten straight losses from a given start is a 0.108% shot. Somewhere in 3,000 rounds it turns up about four times in five.
| Losses the balance can cover | Balance needed | Chance a run goes bust | Runs per bust (about) | Average result per run |
|---|---|---|---|---|
| 6 | 63 units | 1.659% | 60 | −0.062 units |
| 8 | 255 units | 0.423% | 236 | −0.083 units |
| 10 | 1,023 units | 0.108% | 927 | −0.105 units |
Own calculation with a 49.5% win chance per round. A run ends at the first win (+1 unit) or after the last covered loss.
Checking a game and spotting fake predictors
Most crash games keep a list of past rounds on screen. Scroll back 1,000 of them and count how many reached x2. If the game pays 99%, you should land between roughly 46% and 53%. One odd sample happens, so pull a second and a third before taking it to support.
Stake fixed its Crash hash chain in advance and posted the seeding in a BitcoinTalk thread, so the next crash point isn't sitting anywhere you or an app could read it. Aviator "predictors", Telegram signal groups and "hacked" bots want a deposit through their own link, or your login and password.
18+. Crash games carry a house edge at every target, and no cash out pattern removes it. Set a loss limit before you start. Help: Responsible gaming. Not affiliated with Stake, Spribe or InOut Games. Names belong to their owners.
Questions
How often does Crash reach 2x?
The RTP divided by 2. At 99% RTP that is 49.5%, at 97% it is 48.5%.
How much do Stake Crash and Limbo pay back?
Both of Stake's game pages give 99% RTP, a 1% edge and a top multiplier of 1,000,000x.
Is there a crash strategy that beats the house edge?
No. Every target has the same expected return, the RTP. Low targets win often and pay little, high targets win rarely and pay a lot.
Do Aviator or Chicken Road predictors work?
No. Outcomes are locked by seeds or a hash chain that players can only check afterwards. Predictor apps and signal groups are scams that usually want a deposit or your login.
Is Limbo the same as Crash?
The odds follow the same rule. Limbo shows the result instantly against a target you set in advance, Crash lets you cash out while the curve is rising.
Sources
- Stake: Crash game page (Wayback, 14 Sep 2026) · current page on stake.com. 99% RTP, 1% house edge, 1,000,000x max win. Live page returns 403 to our fetcher. Checked 12 November 2026
- Stake: Limbo game page (Wayback, 23 Oct 2026) · current page on stake.com. 99% RTP, 1% house edge, 1,000,000x max Checked 12 November 2026
- Stake: provably fair game events (Wayback, 19 Jul 2025). Limbo formula with houseEdge 0.99, rounded down to two decimals, minimum 1; Crash refers to the seeding thread Checked 12 November 2026
- BitcoinTalk: Stake Crash seeding thread. salted hash chain announced before play, linked from the game events page Checked 12 November 2026
18+ · Play responsibly · gamblingtherapy.org