SSL Secured·✦ Provably Fair·Free Keno result verifier — 100% client-side·18+ Gamble Responsibly
Re-Run The Draw · Private By Construction

KENO PROVABLY FAIR VERIFIER

Ten balls out of forty, allegedly random — here is where you make the casino show its work. From a revealed seed pair and nonce this page regenerates the complete Keno draw so you can hold it against the grid you were paid on.

Enter Your Game Details

Every mode first checks the commitment: SHA-256(server seed) must equal the hashed seed you were shown before betting. Stake and Shuffle seed games use HMAC-SHA256(server seed, "clientSeed:nonce:cursor") per their published docs; BC.Game modes are ported from its open-source verifier; Stake and BC.Game crash use the game hash and salt. For Roobet and other casinos we could not find a byte-level spec, so only the commitment is checked. Sources are listed in the supported list.

Hash
Click to copy
Game Result

Calculation Steps

Keno feels like a lottery, and lotteries have a long, well-documented history of being fixed — which is exactly why crypto Keno’s design matters. Here the drawing machine is an HMAC you can operate yourself: the ten hits are a deterministic function of a committed server seed, your client seed and the bet’s nonce. This page is that machine, outside the casino’s control.

Not sure which to choose? See what we recommend for each type of player.

HOW KENO DRAWS ARE CALCULATED

The tool first checks that SHA-256(revealed server seed) matches the hashed seed you were shown. Then Keno draws 10 distinct numbers from 1–40. Stake (stake.com/provably-fair/game-events) uses a shrinking pool, the same method as Mines:

  1. Generate one float per draw from the HMAC stream.
  2. Start with a pool of [1..40]. For each of 10 draws, idx = floor(float × remaining-pool-size), take that number, remove it from the pool.

Calibration round: the reference seeds some server seed / some client seed at nonce 1 pull the ten numbers 7, 12, 14, 19, 22, 24, 29, 31, 33, 37 (listed sorted here; the tool shows them in draw order).

Shuffle (help.shuffle.com 9286069) maps each float with ceil(float × (40 − i)) and then skips numbers already drawn. Its own worked example, server seed def, client seed abc, nonce 1, draws 18, 27, 7, 24, 35, 11, 28, 9, 31, 16, and the tool reproduces it exactly. Other casinos’ Keno is not reproduced.

HOW OFTEN SHOULD YOUR PICKS OVERLAP THE DRAW?

Keno disappointment is mostly mis-set expectations, and the hypergeometric distribution sets them straight. Mark 10 numbers on the 40-spot card and the draw of 10 overlaps your picks 2.5 times on average — two or three hits is the normal outcome, not a cold streak. Blanking entirely happens about 3.5% of the time (roughly one card in 28), so the occasional zero-hit round is expected behaviour, not evidence of a hostile draw. And sweeping all ten? One card in 847,660,528 — the number of ways to choose 10 from 40. If your hit counts over a session wander around 2–3, the draw is behaving exactly as an honest one must.

WHAT RE-DRAWING A ROUND DEMONSTRATES

A regenerated draw closes the one gap statistics cannot: it ties your specific round to data committed before you bet. The ten numbers this page produces come from a server seed whose hash the casino displayed in advance, mixed with a client seed only you controlled — so if they match the casino’s grid, that draw could not have been picked to dodge your card. Run the numbers on payout value separately with the Keno calculator (paytables differ far more between casinos than the draws do), and see crypto Keno odds for how the math compares across sites.

Related Resources

Play Keno: Crypto Keno Odds

Strategy: Lowest House-Edge Games

Calculate the odds: Keno Calculator · Plinko Calculator · Roulette Calculator

Where the draw is always auditable: fairness-first casino rankings

KENO VERIFIER FAQ

Stake and Shuffle Keno draw 10 numbers from a pool of 40. Your payout depends on how many of your selected numbers appear in those 10.
No. The 10 numbers depend on the server seed revealed only after rotation, so no tool can know them in advance.
In draw order, as the float stream produces them. Casinos usually show the result as hits on the grid, so compare the ten numbers as a set.
For Stake and Shuffle, yes: the tool follows each one's published method and reproduces Shuffle's own worked example. Pick the right casino, enter the revealed server seed, and check that it matches the hashed seed you were shown. Other casinos are not reproduced.