Best Keno Casinos

A keno round has three numbers in it. The size of the board, how many numbers the game draws, and how many you pick. Take a board of 40 numbers where the game draws 10, and call your picks spots. The chance that exactly m of your spots are hit is the count of ways that can happen divided by the count of all possible draws. In the usual notation it is C(k,m) multiplied by C(N-k, n-m), divided by C(N,n), where N is the board, n is the draw and k is your spot count.

That looks heavier than it is. On the 40-number board above, a single spot is hit 10 times in 40, which is 25%. With two spots, both are hit 5.77% of the time, exactly one is hit 38.46% of the time, and neither is hit 55.77% of the time. With three spots, all three come in 1.21% of the time, which is about one round in 82. With five spots, all five come in 0.038% of the time, about one round in 2,611. With ten spots, all ten come in once in 847,660,528 rounds.

One shortcut is worth carrying. On that board you hit a quarter of your spots on average, whatever your spot count is, because the draw covers a quarter of the numbers. Ten spots average two and a half hits. That average is the same for every configuration, which is why the paytable rather than the board decides what you are paid.

Board size and draw count vary between operators. Check both in the game rules before applying any figure from anywhere, including this page. The keno game page covers the variants in circulation.

In games like mines or limbo the operator applies the edge once, as a multiplier on a fair price, and the return is the same everywhere in the game. Keno does not work that way. The payout is a list of fixed figures, one for each number of hits, and the operator writes that list. The return follows from the list rather than from a setting.

The calculation is a weighted sum. Multiply the probability of each hit count by what that hit count pays, then add the results. That total is the return to player. Everything not returned is the house edge.

Work a short one. On a 40-number board drawing 10, a three-spot ticket hits two numbers 13.66% of the time and all three 1.21% of the time. Suppose that table pays 3x for two hits and 46x for three. The return is 0.1366 multiplied by 3, plus 0.0121 multiplied by 46, which comes to about 0.966. That table costs roughly 3.4% of everything staked. Shift the three-hit payout from 46x down to 40x and the return falls to about 89.6%, a cost of 10.4%. One line of the table moved, and the price of the game roughly tripled.

This is the reason keno deserves more scrutiny than a game with a single published RTP. Two casinos can offer an identical board and an identical draw, and one can be three times more expensive than the other without anything on screen looking different. Lottery-operated keno has historically been among the costlier games available anywhere, while in-house versions at crypto casinos are usually far cheaper. The only figure that settles it is the published one for the exact table you are playing.

Most online keno offers a risk setting alongside the spot count, usually named something like classic, low, medium and high. The setting swaps one paytable for another. A low table pays small amounts for modest hit counts. A high table pays nothing at all until you reach a threshold, then pays a great deal.

The effect on your session is large and easy to underestimate. Stay with the 40-number board drawing 10 and pick ten spots. You reach three or more hits about 48.5% of the time, so a table that starts paying at three hits pays on roughly half your rounds. A table that starts paying at five hits pays on about 5% of rounds. The same ticket, the same board, the same average of two and a half hits, and one version returns something twenty times less often than the other.

Operators often hold the return roughly constant across risk settings, so the high table compensates with much larger top payouts. Often is not always, and the arithmetic in the previous section is how you check rather than assume. Run each table you are offered through the weighted sum and compare the totals.

Spot count deserves the same treatment. There is no rule that a five-spot ticket and a ten-spot ticket carry the same return, because they are priced by separate rows of separate tables. Some operators publish an RTP for every spot count. Where they do, read the whole list rather than the headline. Where they publish one figure with no mention of spot counts, treat it as applying to a configuration you have not identified. Our RTP simulator shows what a given return does to a bankroll over a realistic number of rounds.

Start with the published return, per spot count and per risk setting. A single RTP figure for a game with dozens of paytables is not enough information to price a ticket. If nothing is published, you can still compute it yourself from the board size, the draw count and the paytable, and the fact that you have to is worth noting about the operator.

Compare it against what else is on the site. Stake Originals run at 99% RTP, a 1% edge. European roulette costs 2.7% and baccarat on the banker 1.06%. A keno table returning 90% costs ten times what a 99% house game costs per unit staked. Nothing about the presentation will tell you that, which is the point.

Check provable fairness. In a correct implementation the draw is derived from a hashed server seed, your client seed and a round nonce, all committed before you choose your spots. Your selection cannot move the draw. Rotate the seed, take the revealed value and reconstruct a few rounds with our fairness verifier.

Check the auto-play controls. Keno rounds are quick, and bet count caps, stop on loss and stop on profit decide your turnover far more than your number selection does.

Check the bonus terms in full. A headline offer means nothing without its wagering requirement, and keno is frequently weighted below 100% or excluded from clearing one. Run the terms through the bonus calculator first. Our ranking process is documented on the methodology page, and the tools on the responsible gambling page are worth switching on before a session rather than during one.