Keno simulator with true hypergeometric odds: see the payout table and hit probability for each number of picks.
Pick k numbers from a grid, the game draws a fixed set, and you are paid by how many of your picks it hit. The probability of each hit count follows the hypergeometric distribution — drawing without replacement, which is why the maths is less friendly than it looks.
The simulator prints the true probability of every hit count next to the payout, so you can see exactly which rows of the table carry the return.
Choosing ten numbers rather than three makes hitting some of them near-certain and hitting most of them vanishingly unlikely. The payout table compensates by paying nothing for low hit counts, so the effective outcome is a longer wait for a bigger, rarer win.
Expected value stays flat across pick counts on a well-built keno; what changes is variance and how often you see anything at all.
Keno odds are hypergeometric: drawing without replacement from a fixed pool. Picking more numbers raises the chance of hitting some of them and collapses the chance of hitting most of them, which is why the payout for a full house grows so steeply.
Across sensible paytables the expected return is roughly constant regardless of how many numbers you pick. The choice is between many small returns and a remote chance of a large one.
No. Each draw is independent and every number is equally likely, whatever the history display suggests.
On house-game keno at 99% RTP, considerably. Lottery-style keno in land-based venues is often far worse.