The draw has no memory, and that is the whole answer
Keno sits in a strange middle ground between a lottery and a casino table game. You pick a set of numbers, the house draws a larger set from the same pool, and you are paid according to how many of your picks match the draw. The question about which numbers come up most often sounds like it should have a data-driven answer, and plenty of websites will sell you one in the form of hot and cold number charts. The honest answer is simpler and less exciting: in a fair draw, every number in the pool has exactly the same chance of appearing on every single game, and past draws change nothing about the next one.
A standard keno game draws 20 numbers from a pool of 80. Before any draw, each of the 80 numbers has a 20-in-80 chance of being among the 20 selected, which is one in four. That probability is identical for the number 7, the number 68, and every number in between. It was identical yesterday, it will be identical tomorrow, and it does not shift because a particular number has appeared in five consecutive draws or has been absent for fifty.
The reason frequency charts exist is that humans are pattern-seeking creatures. If you record a few hundred draws, some numbers will have appeared slightly more often than others, just as flipping a coin a hundred times rarely produces exactly fifty heads and fifty tails. Those differences are what statisticians call sampling noise. They are not evidence that the game favours certain numbers, and they have no predictive power for the next draw.
Why hot and cold number lists look convincing
Keno operators and third-party sites sometimes publish lists of the most frequently drawn numbers over a recent period. A player sees that 23 has come up 27 times in the last 100 games while 41 has come up only 19 times, and the conclusion feels obvious: 23 is hot, 41 is cold, and the smart play is to bet the hot number because the wheel has no memory. The arithmetic behind that feeling is worth unpacking, because it is the same arithmetic that powers every lottery and keno myth.
With 20 numbers drawn from 80, the expected count for any single number over 100 games is 25 appearances. The actual count will scatter around that average. Some numbers will land at 21, others at 29, a few further out. That scatter is exactly what a random process produces. If every number sat at precisely 25 appearances over any 100-game window, the draw would not be random; it would be engineered to look balanced. The unevenness is the fingerprint of randomness, not a flaw in it.
Keno draws are independent events. The machine or the random number generator does not consult a ledger of recent results before selecting the next 20 numbers. A number that has appeared often has the same one-in-four chance on the next draw as a number that has appeared rarely. The hot number is not more likely to continue, and the cold number is not due to catch up. Both beliefs are versions of the gambler's fallacy, and both quietly hand the player a reason to make bets they would not otherwise make.
What varies in a keno game
None of this means every keno ticket is identical. The choices that matter are not which numbers you pick, but how many you pick and how the pay table rewards a match. Those two factors interact in a way that genuinely changes the expected rate of loss, which is the only number in the game that is stable enough to plan around.
Keno pay tables are built so that the house keeps a mathematical advantage on every ticket. That advantage is typically larger than the edge on most table games, and it varies by how many numbers you select. A one-spot ticket pays a fixed amount if your single number is drawn, and the house edge on that ticket is a specific number printed in the game's pay table. A ten-spot ticket pays enormous odds for catching all ten, but the probability of doing so is vanishingly small, and the intermediate payouts are set so that the overall return still favours the house.
The table below lays out the real decision points in keno and what each one does and does not change.
| Decision | What it changes | What it does not change |
|---|---|---|
| Which numbers you pick | Nothing about the probability of a match | The draw probability is identical for every number |
| How many numbers you pick | The pay table you are playing against, and therefore the house edge | The randomness of the draw itself |
| Whether you use a hot or cold number list | Only your confidence in a pattern that is not there | The independence of each draw |
| The stake per ticket | The absolute size of wins and losses | The percentage edge built into the pay table |
| The specific pay table offered | The expected return for the same number of spots | The underlying draw mechanics |
The practical takeaway is that a player who wants the least punishing keno ticket should compare pay tables for the same spot count across the games available to them, not hunt for lucky numbers. The pay table is the one part of the game that is printed, fixed, and knowable in advance. The draw is none of those things.
How the draw is generated, and why it matters for this question
In a physical keno game, the numbers are selected by a blower machine that mixes balls and draws them, or by an equivalent mechanical process. In an online or video keno game, the selection is produced by a random number generator. Both methods are designed to produce a uniform distribution, meaning every number in the pool has the same chance of being selected on every draw.
Independent testing laboratories commonly certify the random number generators used in regulated online games. That certification covers whether the generator produces outcomes that are statistically indistinguishable from a fair random draw. It does not cover an operator's financial conduct, and it is not a promise about any individual game. But it is the reason a player can reasonably treat the draw as fair rather than as a process with hidden favourites.
Some crypto-based keno games add a provably fair layer, where a hashed server seed is published before play and revealed afterwards. That lets a player verify that the outcome of a particular round was not altered after their bet was placed. It verifies individual round integrity. It does not change the underlying probability, and it does not make any number more or less likely to appear. A provably fair keno game still draws 20 numbers from 80 with a uniform distribution, because that is what the game is.
What a frequency chart can and cannot tell you
If you have access to a history of past draws, you can compute exactly how often each number has appeared in that history. That computation is real, and the resulting chart is accurate as a description of the past. The error is in treating the chart as a forecast.
Suppose a number appeared 30 times in the last 100 draws. The chart tells you that fact and nothing more. It does not tell you that the number is more likely to appear in draw 101. It does not tell you that the number is less likely to appear, either. The next draw is a fresh event with the same one-in-four probability for every number. A chart that implies otherwise is selling a story, not a strategy.
The same logic applies to patterns of pairs or sequences. A player might notice that 12 and 47 have appeared together unusually often and conclude that the pair is somehow linked. In a random process, every pair of numbers will show some degree of co-occurrence over a finite sample, and some pairs will show more than others by chance alone. There is no mechanism in a fair keno draw that would cause two numbers to attract or repel each other.
Why the question usually hides a different hope
Most people asking which keno numbers come up most often are not really asking for a frequency table. They are asking whether there is a way to pick numbers that improves the chance of winning. The honest answer is that there is not, and the reason is structural rather than a matter of insufficient data.
Keno is a negative-expectation game. The pay table is constructed so that, over a very large number of tickets, the player receives back less than the total amount wagered. No selection method changes that. Picking birthdays, picking hot numbers, picking cold numbers, picking the same numbers every time, picking a random quick pick: all of these produce the same expected return for a given spot count and pay table, because the draw does not care how the numbers were chosen.
What a player can control is the size of the built-in disadvantage. Comparing pay tables, playing fewer spots when the pay table is less generous, and treating keno as an entertainment expense with a known expected cost are all legitimate choices. They are choices about how much the game costs to play, not choices about how to beat it. No betting pattern, number selection system, or progression of stakes changes the edge that is printed into the pay table.
FAQ
Do some keno numbers really hit more often than others?
Over any finite set of past draws, some numbers will have appeared more often than others by chance alone. That does not mean those numbers are more likely to appear in future draws. In a fair game, every number has the same one-in-four chance on every draw, and past frequency has no predictive value.
Is it better to play the same keno numbers every time?
Playing the same numbers every time does not change the probability of winning. Each draw is independent, so a set of numbers that has lost many times is not any closer to hitting. The only thing that changes your expected return is the pay table and how many spots you choose, not which numbers you choose or whether you stick with them.
How do I find the keno pay table before I play?
The pay table is normally shown on the game screen or in the game's information panel, often behind a button labelled something like pay table, info, or help. It lists the payout for each number of matches at the spot count you have selected. Comparing that table across different keno games is the most reliable way to see which game charges the smallest built-in edge.
Gambling involves risk. Please only gamble with funds that you can comfortably afford to lose. While we do our best to offer good advice and information, we cannot be held responsible for any loss that may be incurred as a result of gambling. We do our best to make sure all the information that we provide on this site is correct. However, from time to time mistakes or changes may occur. Some offers mentioned on this page may no longer be available, or may have changed. We encourage you to check the terms and conditions of any offer before you take advantage of it. If you or someone you know is struggling with gambling addiction, please seek help. There are many organizations that can provide support and advice, such as GamCare, Gamblers Anonymous, or the National Council on Problem Gambling. In the UK, you can also contact the National Gambling Helpline. Remember, gambling should always be a form of entertainment, not a way to make money. Please play responsibly.
Play responsibly
Gambling should be treated as paid entertainment, never as a way to earn income or recover losses. 1-800-MY-RESET (1-800-697-3738)
18+ or 21+ depending on where you are; follow the minimum age that applies to you.
This article is general information about how these mechanics work. It is not legal advice and not a recommendation to gamble or to use any particular operator. Availability and legality differ by jurisdiction — check the rules that apply where you are.