Keno Strategy Guide: Odds, Number Picking and Payouts
Keno strategy cannot make one number more likely than another because a standard game draws 20 numbers from 1 through 80 without replacement. The practical decisions are choosing a pay table, understanding spot-count variance, limiting total action, and avoiding the belief that hot, cold, overdue, or patterned numbers predict the next independent draw.
Standard Keno Rules
A player selects a group of numbers, usually called spots. The game draws 20 of the 80 available numbers. Payouts depend on:
- how many spots were selected;
- how many selected numbers were drawn;
- stake per ticket;
- the exact pay table;
- whether a multiplier or bonus feature applies.
A 10-spot ticket chooses ten numbers. If five of those ten appear among the twenty drawn, the result is five catches or five hits.
Different casinos can use the same 80-number and 20-draw structure while offering very different returns. A 5-spot table at one venue is not interchangeable with another.
Hypergeometric Probability
Keno uses sampling without replacement. The probability of hitting exactly k numbers after selecting n spots is:
P(k hits) = C(n,k) × C(80-n,20-k) ÷ C(80,20)
C(a,b) is the number of ways to choose b items from a without considering order.
The denominator counts all possible 20-number draws from 80 numbers. The numerator chooses k hits from the selected spots and fills the remaining draw positions from unselected numbers.
One-Spot Keno
A single chosen number appears in 20 of the 80 draw positions:
P(hit) = 20 ÷ 80 = 25%
The miss probability is 75%. If a one-spot ticket pays 3-for-1 as a total return on a hit, expected return is:
0.25 × 3 + 0.75 × 0 = 0.75 units
The house edge is 25%. If the wording is 3-to-1 profit plus the returned stake, the total return is four units and EV is break-even before other rules. This distinction between for-1 and to-1 is essential.
Five-Spot Probabilities
For five selected numbers, the exact standard probabilities are approximately:
| Hits | Probability | Approximate frequency |
|---|---|---|
| 0 | 22.7184% | 1 in 4.40 |
| 1 | 40.5686% | 1 in 2.46 |
| 2 | 27.0457% | 1 in 3.70 |
| 3 | 8.3935% | 1 in 11.91 |
| 4 | 1.2092% | 1 in 82.70 |
| 5 | 0.06449% | 1 in 1,550.57 |
The probabilities follow directly from the hypergeometric formula. Number choice does not alter them.
Five-spot EV example
Consider a hypothetical table where total returns per unit are:
| Hits | Total return |
|---|---|
| 0 or 1 | 0 |
| 2 | 1 |
| 3 | 2 |
| 4 | 15 |
| 5 | 500 |
Expected return is:
(0.270457 × 1) + (0.083935 × 2) + (0.012092 × 15) + (0.0006449 × 500)
Expected return = 0.942175 units
The house edge is:
1 - 0.942175 = 0.057825, or about 5.78%.
At $500 of total action, projected loss is:
$500 × 0.057825 = $28.91
This is a pay-table demonstration, not a claim about a live casino game.
Ten-Spot Probabilities
A 10-spot ticket produces the following selected probabilities:
| Hits | Probability |
|---|---|
| 0 | 4.5791% |
| 1 | 17.9571% |
| 2 | 29.5257% |
| 3 | 26.7402% |
| 4 | 14.7319% |
| 5 | 5.1428% |
| 6 | 1.1479% |
| 7 | 0.1611% |
| 8 | 0.01354% |
| 9 | 0.000612% |
| 10 | 0.00001122% |
Hitting all ten occurs with probability about 0.0000001122, equivalent to roughly 1 in 8,911,711.
Exactly five hits is much more common at approximately 5.1428%, or about 1 in 19.45. A pay table can therefore return a modest amount for five while reserving a large top prize for ten.
The headline jackpot does not reveal expected return. Every payout tier and probability must be included.
Does Picking More Numbers Improve the Odds
Selecting more spots increases the chance of hitting at least one number because more numbers are covered. It also creates harder top-hit targets and a different payout distribution.
For a 20-spot ticket, the most likely outcome is five hits at about 23.33%. Hitting all 20 has probability around 2.83 × 10⁻¹⁹, or roughly 1 in 3.54 × 10¹⁸.
The useful comparison is not hit frequency alone. It is:
Expected return = Σ(probability of tier × total return for tier)
A game can hit frequently while returning less than the stake on many winning tickets. A lower-hit-frequency ticket can have a better RTP if its pay table is stronger.
Number Picking Systems
Hot numbers
A hot-number display shows numbers drawn frequently in a recent sample. If each draw is independent and the equipment is fair, past frequency does not increase the next-draw probability.
Every specific number still has a 25% chance of appearing in a standard 20-of-80 draw.
Cold and overdue numbers
A number missing from ten previous draws is not due. The chance it misses the next draw remains 75% under the independent model.
The probability that one fixed number misses ten consecutive draws is:
0.75¹⁰ ≈ 5.63%
Such droughts occur naturally across 80 numbers.
Birthdays and visual patterns
Choosing birthdays concentrates selections from 1 through 31 but does not change mathematical probability. Lines, diagonals, clusters, lucky numbers, and random quick picks also have equal draw probabilities.
Pattern choice can affect jackpot sharing only when a prize is divided among multiple winning tickets and many players choose the same combinations. Most fixed-payout keno tables do not use a shared jackpot for ordinary tiers, so the practical effect depends on rules.
Repeating the same ticket
Repeating numbers gives the same combination multiple independent opportunities. Changing numbers also gives multiple opportunities with equal probability. Consistency can simplify recordkeeping but does not increase per-draw EV.
Comparing Pay Tables
A proper comparison records every tier as a total return per unit. Pay tables may show:
- payout including returned stake;
- net profit with stake returned separately;
- payouts only at maximum bet;
- capped aggregate liability;
- progressive jackpots;
- different payouts by denomination;
- bonus-ball or multiplier results.
For each spot count:
- calculate or obtain every hit probability;
- convert every award to total return per unit;
- multiply probability by return;
- sum the contributions;
- subtract the result from one for house edge.
If a progressive applies, include the current jackpot and probability of winning it. A meter can improve RTP as it grows, but taxes, shared prizes, caps, and eligibility conditions matter.
Ticket Cost and Total Action
Keno often runs quickly, especially online. Total action is:
Ticket cost × tickets per draw × number of draws
Two $2 tickets per draw over 100 draws create:
$2 × 2 × 100 = $400 action
At a 12% house edge, projected loss is:
$400 × 0.12 = $48
Doubling the number of tickets doubles expected dollar loss when edge and stake remain constant. It does not diversify risk in the investment sense because all tickets retain negative expectation.
Variance and Long Losing Runs
A pay table with a large top prize can have high variance. The theoretical RTP may depend heavily on an event occurring once in thousands or millions of tickets.
For a binary one-spot game paying three total units with 25% probability, expected return is 0.75 and variance of return is:
Variance = 0.25 × (3 - 0.75)² + 0.75 × (0 - 0.75)²
Variance = 1.6875, giving a standard deviation of about 1.299 units per ticket.
Over 100 independent tickets, standard deviation is approximately:
1.299 × √100 = 12.99 units
The projected loss is 25 units, but realized results can vary widely. High-spot games with extreme jackpots usually have even larger variance.
Bankroll and Risk Limits
Keno bankroll management cannot remove the house edge. It can limit exposure.
A practical structure is:
- use a fixed ticket cost no larger than 0.5% to 1% of the session bankroll;
- set a maximum number of draws;
- cap total tickets per draw;
- record action rather than only wins and losses;
- avoid adding spots or multipliers to chase a loss;
- choose a pay table before choosing lucky numbers;
- avoid borrowed or essential money;
- stop when the precommitted limit is reached.
With a $200 session bankroll, a 1% ticket unit is $2. A 50-draw limit creates at most $100 action for one ticket per draw. Buying five tickets per draw would exhaust the same action cap in ten draws.
RNG, Ball Draws, and Fairness
Online keno usually uses a certified random-number generator. Retail or casino versions may use electronic systems, video displays, or ball draws. The game rules should explain the selection process and audit framework.
A previous draw should not influence the next RNG result. Duplicate draw histories, unexplained result changes, or missing rules are reasons to stop and use the operator's dispute process.
Legal eligibility also matters. A technically accessible keno site is not necessarily licensed for the player's location. Identity, age, geolocation, and account rules should be followed without VPNs or false information.
Keno Strategy Checklist
| Decision | Best analytical approach |
|---|---|
| Spot count | Compare full pay-table RTP and variance |
| Number selection | Treat all combinations as equally likely |
| Ticket frequency | Set a fixed action cap |
| Multiplier | Calculate the additional cost and weighted payouts |
| Progressive | Include current meter and win probability |
| Session length | Precommit to draws and time |
| Bankroll | Use small fixed units and no chasing |
| Online game | Verify rules, license, RNG, and location eligibility |
Conclusion
Keno has no predictive number-picking system. Hot, cold, overdue, patterned, and repeated selections have the same next-draw probability in a fair 20-of-80 game. Useful strategy consists of calculating hypergeometric odds, comparing the complete pay table, choosing a tolerable variance level, and limiting tickets and draws. These controls cannot guarantee profit, and every large prize remains a rare event rather than a recovery plan.
Guide last updated: July 2026. GamblersScore covers keno strategy for adults who want probability formulas, payout EV, and clear bankroll limits.