How Keno Odds Are Calculated: The Hypergeometric Basis
Keno uses a fixed pool (typically 80 numbers) from which the house draws a fixed number of winning spots (commonly 20). If you select r numbers (called “spots”), the number of hits you get in a single draw follows a hypergeometric distribution: there are N total items (N = 80), K success states in the population (K = 20 drawn numbers), n draws from the population (n = r numbers you picked), and the probability of exactly h hits is given by C(r,h) * C(80-r, 20-h) / C(80,20). A simpler way to see specific probabilities is via sequential product terms: the probability that all r chosen numbers are among the 20 drawn equals (20/80) * (19/79) * ... * ((20-(r-1)) / (80-(r-1))). For r = 1 that reduces to 20/80 = 0.25. For r = 2, the probability both hits occur is (20/80)*(19/79) ≈ 0.0601. This product form is useful for computing the probability of a full-house (all spots hit) without computing large combinatorial numbers. For other exact-hit counts (e.g., exactly h hits), the hypergeometric formula is the correct tool and can be computed directly or via statistical software. Understanding this combinatorial backbone clarifies why larger spot plays are exponentially harder to complete perfectly and why the game's payouts must reflect these steeply diminishing probabilities.
Understanding Hit Frequencies, Expected Hits, and Variance
Although players often focus on jackpot-type hits (hitting all chosen spots), the more relevant long-run metric is the expected number of hits and the distribution around that expectation. The expectation for hits when choosing r spots is r * (20/80) = r * 0.25, because each selected spot independently has a 25% chance to appear among the 20 draws (more precisely, hypergeometrically dependent, but the linearity of expectation still holds). So if you play 4 spots on many independent draws, the long-run average hits per ticket will be about 1.0. Variance quantifies volatility: hypergeometric variance is r*(K/N)*(1-K/N)*((N-r)/(N-1)), which simplifies in keno to r*0.25*0.75*((80-r)/79). For small r, variance is modest; for larger r variance grows, reflecting wild swings—many small wins vs. occasional larger wins. Standard deviation is the square root of that variance and tells you how much results will typically fluctuate around the mean. Practically, this means short sessions may deviate substantially from expected hits, so your observed win/loss over 50–100 draws can be far from the long-run averages. Because draws are without memory and independent from past outcomes, the only reliable statements are probabilistic. The hypergeometric nature also explains why "due" or "hot/cold" number beliefs are fallacious: the distribution of hits each draw is governed only by combinatorics, not by prior draws. For players, knowing the expected hits helps align realistic expectations (e.g., how often you might expect a single or double hit on 4-spot plays) and to size bets and session lengths accordingly.
How Payout Tables Translate Probability into House Edge
A keno payout table maps each possible hit count to a reward for a given stake. The expected return (or expected value, EV) of a single ticket equals the sum over all hit counts of [probability of that hit count × payout for that hit count] minus the cost of the ticket. House edge is derived from EV: House Edge = (Wager − Expected Return) / Wager. For example, if a $1 ticket returns on average $0.70, the house edge is 30% (the player loses 30 cents on average per dollar bet). Different casinos and different keno variants provide widely differing pay tables; some 1-spot or 4-spot tables have comparatively high return-to-player (RTP), while other spot counts may be heavily unfavorable. Importantly, identical probability distributions can produce drastically different house edges if the payout schedule changes. A common observation is that short-spot games (1–4 spots) often have higher RTPs relative to long-spot games, because payouts for hitting many numbers have to be astronomically large to compensate for tiny probabilities, and most establishments pay less than mathematically fair amounts on those rare events. To compute expected return yourself, tabulate probabilities for each hit count using the hypergeometric formula, multiply each by the associated payout, sum the results and subtract your wager. That gives you the long-run average return; dividing by the wager and subtracting from 1 yields the house edge. Transparent players compare pay tables across venues (including online operators) and choose games or spot counts with the most favorable EV if they aim to maximize longevity of play and minimize expected loss.
Practical Play Strategies: Spot Selection, Bankroll, and Risk Management
No strategy can overcome a negative expected value offered by the house, but understanding odds helps optimize play for entertainment value and risk tolerance. Spot selection influences both hit expectation and volatility: fewer spots (1–3) provide frequent modest wins and lower variance; mid-range spots (4–6) often have interesting jackpots and sometimes the best published RTPs depending on the pay table; high spot counts (7–20) produce rare large payouts but extremely long losing streaks. Practical tactics include flat betting (wagering the same small amount each ticket) to smooth variance, setting session loss and win limits to avoid chasing losses or giving back gains, and using unit stake sizing so you can sustain many draws — for example, maintaining a bankroll that allows at least 50–200 average-ticket plays depending on your risk appetite. If you play for the thrill of big payouts, accept the higher house edge and treat the cost as entertainment; if you care about minimizing expected loss per hour, choose games and spot counts with the highest RTP on the pay table and keep bets modest. Some players use coverage systems (buying multiple tickets to cover more numbers across a draw) or wheeling strategies to increase the chance of partial hits, but these simply change the payout distribution and rarely improve EV once ticket cost is considered. Lastly, use available tools: calculators and simple simulations (Monte Carlo) can illustrate long-run outcomes for your chosen spots and pay tables so you can make an informed decision before staking real money.
