Overview
Keno: how the draw works and how to count the odds
Keno is a number-draw game. You mark some numbers, a set of balls is drawn, and the result depends on how many of your numbers were drawn. This overview explains the model behind every keno guide on this site.
The game in brief
A keno board holds N numbers. Before the draw you mark k of them; this count is your pick count or number of spots. The game then draws D numbers at random without replacement, so no number can appear twice. Your result is the number of hits, meaning how many of your k numbers are among the D drawn. A pay schedule says what each hit count returns. Nothing else matters: not where the numbers sit on the grid, not which ones were drawn last time.
The two common formats
Online keno services often use a board of 40 numbers with 10 drawn per game. The traditional format, found in lottery-style keno, uses 80 numbers with 20 drawn. Both draw exactly one quarter of the pool, so a single marked number has the same chance of 25.00% on either board. Once you mark two or more numbers, the chances begin to differ, because removing one ball from a pool of 40 changes the remaining odds more than removing one from 80.
The hypergeometric model
Counting hits is a classic sampling problem. Think of the board as N balls, of which D are "drawn" and N - D are not. Your ticket is a sample of k balls. The chance that exactly h of them are drawn balls is:
P(h hits) = C(D, h) x C(N - D, k - h) / C(N, k)
Here C(n, r) is the number of ways to choose r items from n, also written "n choose r". The top of the fraction counts tickets with exactly h drawn numbers and k - h undrawn ones. The bottom counts all possible tickets. For example, on the traditional board with five spots and three hits: C(20, 3) = 1,140, C(60, 2) = 1,770, and C(80, 5) = 24,040,016. The chance is 1,140 x 1,770 / 24,040,016 = 8.39%.
Two quick facts follow from the model. The expected number of hits is k x D / N, so five spots on 80/20 average 1.25 hits. And the probabilities for all h from 0 to k always add to exactly 1, which is a useful check on any table you build.
| Hits | Probability | As odds |
|---|---|---|
| 5 | 0.0645% | 1 in 1,550.6 |
| 4 | 1.2092% | 1 in 82.7 |
| 3 | 8.3935% | 1 in 11.9 |
| 2 | 27.0457% | 1 in 3.7 |
| 1 | 40.5686% | 1 in 2.5 |
| 0 | 22.7184% | 1 in 4.4 |
From probabilities to a return figure
A pay schedule turns the table above into money. Multiply each hit count's pay by its probability and add the results: that sum is the return to player per unit staked. If it is below 1, the difference is the house edge, and the ticket loses that share of stakes over time. The pick count does not escape this; it only changes how the losses and the occasional wins are spread out. Schedules on this site are always labelled hypothetical; real services publish their own.
What the keno guides cover
Each keno guide on this site takes one question and works it through with exact numbers. The guides grouped by pick count derive the full hit distribution for one ticket size, with at least one line counted by hand:
- 40/10 board, one page per pick count from 1 to 10, starting with the one-number ticket and ending with the ten-pick ticket.
- 80/20 board at two, four, six and eight spots, from the two-spot page to the eight-spot page.
- A ten-spot comparison of the two boards line by line.
- Other guides deal with pay table returns, hot and cold number myths, hash-based draws, simulations, streaks, bankroll limits, variance and a glossary.
For the hash method that lets you verify a draw, read the hash fairness guide. For how a house edge is defined across game types, see the house edge guide.