AviatorAviator simulator & provably-fair round verifier
Two genuine, math-correct tools that run entirely in your browser. The simulator proves no cashout target beats the house edge; the verifier shows how a finished round is checked — the reason prediction beforehand is impossible.
Set a target multiplier and run thousands of provably-fair rounds. Watch your realised RTP converge toward the house figure — proving no target beats the edge.
Bankroll equity curve (dashed line = starting bankroll):
Last 50 rounds:
| # | Multiplier | Result | Bankroll |
|---|
Paste the seeds a casino reveals after a round to independently recompute it. This works only post-round — which is exactly why prediction beforehand is impossible.
Why a multiplier mismatch is usually not cheating
The exact byte order, text encoding and the precise hash-to-multiplier conversion can differ slightly between Spribe's live implementation and the public model used here. If the SHA-256 commitment matches but the multiplier differs, the cause is almost always formatting, rounding or edge-handling — not a tampered result. The commitment check is the security-critical one: it proves the server seed was fixed before the round.
The Aviator probability table
The chance of the multiplier reaching at least a target X is simply P = 0.97 ÷ X. Higher targets pay more but hit less often — and the two effects cancel, so expected value stays at −3% everywhere.
| Cashout target X | Chance of reaching X (RTP 0.97) | Rough frequency |
|---|---|---|
| 1.5× | 64.7% | Very frequent; ~1 in 1.5 rounds |
| 2× | 48.5% | Roughly a coin-flip |
| 3× | 32.3% | About 1 in 3 |
| 5× | 19.4% | About 1 in 5 |
| 10× | 9.7% | About 1 in 10 |
| 50× | 1.94% | About 1 in 52 |
| 100× | 0.97% | About 1 in 103 |
How to read this
Around 70% of rounds reach roughly 1.5× and the great majority stay modest, which is why low-target “grinding” feels safe — but the long-run edge is identical to swinging for 100×. Use the table to understand variance and session feel, never as a promise of profit. Want the full derivation? See our RTP & strategy guide.
How each tool works
Simulator: the round engine
For each round we draw u in (0,1] and set the crash multiplier m = RTP / u, floored to two decimals with a 1.00 minimum. If m ≥ your target you win bet × X; otherwise you lose the bet. An optional second bet settles independently against the same m, modelling the dual-bet hedge. We then report hit rate vs theory, realised RTP, streaks, the longest drought, and an equity curve.
Verifier: the two checks
Commitment: we hash the revealed server seed with SHA-256 and compare it to the pre-round hash the operator published. Result: we concatenate the server seed and the first three client seeds and nonce, hash with SHA-512, take the first 13 hex characters as h, and derive the multiplier with the standard crash formula. Both run via the browser's Web Crypto API, fully transparently.
Tool FAQ
Is the simulator using real Aviator data?
It does not connect to any live game — that would be impossible, because results do not exist until a round runs. Instead it reproduces the exact crash maths: each round draws a uniform random number u and computes the multiplier as RTP ÷ u (floored to two decimals, minimum 1.00). This yields the correct distribution where the chance of reaching X equals RTP ÷ X. Over many rounds your realised return converges to the RTP you set, which is the whole point.
Why does every target give the same long-run return?
Expected value per round = P(win) × (X − 1) − P(lose), and with P(win) = RTP ÷ X the multiplier algebraically cancels, leaving EV = RTP − 1 = the negative house edge. A higher target pays more but hits less often by exactly the offsetting amount. The simulator makes this visible: try 1.5×, 2× and 10× over 100,000 rounds and all three converge to roughly the same RTP.
Does the verifier prove a casino cheated if the multiplier does not match?
Not on its own. The security-critical test is the SHA-256 commitment: if the revealed server seed hashes to the value the operator published before the round, the seed was fixed in advance and could not have been swapped. A multiplier mismatch with a passing commitment is almost always a formatting, encoding or rounding difference between the public model and the operator's exact implementation, not tampering.
Can I use these tools to win?
No, and that is deliberate. The simulator exists to show that no auto-cashout target and no "system" beats the −3% edge, and the verifier exists to show prediction is impossible because results are only knowable after the round. They are honesty tools, not winning tools. Both run entirely in your browser, with no sign-up and no real money.