The one game we didn't have to sample
Unlike a slot, Keno hides nothing: it's a grid of 40 numbers. You pick up to 10, the game draws 10, and your payout depends on how many of your picks land based on a paytable. So the odds are exact, computable, and identical at every casino (pure hypergeometric probability). We didn't need to play thousands of rounds; we calculated it.
The only thing that changes between casinos or risk levels is that paytable, which is the multiplier for each number of hits. Those tables aren't published anywhere (which is why no explainer for the risk levels exists). We used the real Stake pick-10 tables, read off the game and cross-checked until our math reproduced Stake's stated RTP to the decimal. Everything below follows from that.
One of the most-played games in crypto
Keno can feel like a sideshow, but the numbers say otherwise. In a single week, players placed 3.8 million bets on it across 25 crypto casinos we track (about 545,000 a day). The play is also wildly concentrated: Gamdom alone took 28%, with the top five sites soaking up nearly three-quarters of it.
First: the two things you actually choose
Before looking at any charts, it helps to understand the core mechanics. Keno gives you two dials you can control:
1. How many numbers you pick (from 1 to 10).
2. The "risk" level (Classic, Low, Medium, or High).
You do not choose how many of your numbers actually hit. The draw decides that automatically.
A common myth is that picking one number gives you a 1-in-40 shot. Because the game draws 10 numbers out of 40, a single pick actually lands 1 in 4 times (25%). However, the more numbers you pick, the harder it is to hit all of them. Here is how those odds scale in a fair game with no house edge:
| You pick | Chance all of them hit | What a fair game would pay |
|---|
This answers a common question: If you usually hit 3 numbers, why not just pick 3?
Because matching 3 numbers is a completely different bet depending on your grid setup:
- Pick 3, Hit 3: A rare 1-in-82 long shot. Matching all three is your jackpot, so it pays big.
- Pick 10, Hit 3: A common result that lands almost every third round (1 in 3.5). Matching three is just a small consolation payout because you were aiming for ten.
In short, picking fewer numbers creates a rare, all-or-nothing bet. Picking more numbers creates frequent, small payouts. The pick count is a volatility dial on its own.
To keep the rest of this study clear, we lock the first dial at the most popular setting of 10 picks. From here, we focus on the second dial: what Low, Classic, Medium, and High risk settings actually do to your money.
You expect two and a half hits. That's it.
Before looking at paytables, let's start with the math behind every round. Pick 10 numbers out of 40, and the draw hands you an average of just 2.5 hits. Landing 2 or 3 is the overwhelmingly likely outcome; 5 is already a 1-in-24 event, and hitting all 10 is 1 in 848 million. No risk setting and no casino can nudge these by a hair because they are fixed by the math. The risk switch only changes what you are paid for each of them.
What your everyday round is worth
Here's the main takeaway: you don't choose how many numbers hit. The draw does, and on about 6 rounds in 10 it lands on 2 or 3 (the average is 2.5). The only lever you pull is the risk level. So the real question is simple: on those everyday 2- and 3-hit rounds, does your risk setting pay you back or not? Here is the answer as your net profit or loss on a $10 bet.
| The risk you choose | If you hit 2 (~31%) | If you hit 3 (~29%) | What that means for you |
|---|
So why does a 2 pay nothing on Classic, Medium and High, while a 3 pays? Because with 10 picks you average 2.5 hits, so hitting 2 is not really a win. Every risk level has the same ~99% to give back; the dial only decides where the payouts switch on.
Low switches on early. It hands you a token 1.10x even for a below-par 2, so you win small and constantly (it just feels like winning, but the edge is identical). Classic and Medium don't waste budget on a below-average 2: they save it to pay more when you beat the average with a 3 (which is why their 3 pays 1.40x and 1.60x versus Low's 1.20x). High waits longest of all, paying nothing until 4, then paying big. Same pot of money, just handed back at different moments: that is the entire difference between the settings.
Same edge, wildly different ride
Here is the answer that isn't written down anywhere. All four risk levels return the same ~99%, so the switch is not an edge choice: it's a volatility choice. Low pays something on 4 rounds in 5 in tiny amounts (a gentle bleed). High pays nothing 4 rounds in 5, but its wins land harder. The payout ladder below shows it: Low starts paying at 2 hits but stays low, while High pays nothing until 4 hits and then climbs steeply.
| Risk level | RTP | Pays something | You get nothing | Volatility (CV) | Top prize |
|---|
Notice two things in that table. All four RTPs sit within a third of a percent of each other, meaning the edge is flat. Also, Classic and Medium are near-identical twins (same 48.5% win rate, near-identical volatility), differing only in a cosmetic top prize. So Stake's four levels are really three: a calm one, a middle pair, and a wild one.
That 1,000x is almost pure decoration
Low, Medium and High all display a 1,000x top prize. Here is what the promo won't tell you: to reach those flashy multipliers you need to hit 8 or more of your 10 numbers (a 1-in-42,647 event). Those giant wins deliver a tiny fraction of the actual payback. Around 99% of everything Keno gives back comes from hitting 5 or fewer. The big number on the box is there to be seen, not to be paid.
100 rounds at $1, four different ways
What the dial actually feels like: here is the spread of where you land after 100 rounds at $1 a go on each risk level. Every level drifts to roughly the same small loss on average (that's the shared ~1% edge), but the spread is the whole story. Low keeps you in a tight band near break-even. High swings from deep losses to big wins: same expected cost, completely different volatility rating.
What 600 rounds actually look like
Now looking at the long game: $100, $1 a round, 600 rounds. Flip between the risk levels. The average ending is nearly identical (a gentle ~1% drift down to about $92), but the ride could not be more different. Low drifts down in a narrow, predictable channel. High is a rollercoaster where lucky players double their money and unlucky ones wipe out. On every setting, ~95% of players are ahead at some point, yet only about 1 in 3 are still ahead at the end. You will almost certainly feel like a winner before the math catches up.
Your casino matters more than your risk
The risk switch moves your edge by less than 0.75%. But which casino you play on moves it by nearly 4%: from Duel's near-perfect 99.95% down to Roobet's ~96%. That is the lever that actually matters. Duel and Gamdom go further with cash-back rebates on top, while BC.Game and Roobet quietly keep two-to-four times as much of every bet.
What the math says before you pick
The confirmed paytables
The full multiplier tables behind this page show what each number of hits pays on every risk level across all three RTP tiers we verified. Toggle between them: Duel's near-perfect table, Stake's 99%, and BC.Game's 98%. Same game, same fixed odds: only these numbers change, and they are the entire difference between a 99.99% site and a 98% one.