Wild.io
Keno
KenoBeginner

Best Keno Numbers to Play: Myth vs Math

Every keno number has the same 25% chance per draw, no matter what the lucky-number lists say. We trace the "1, 4, 23, 34, 72" meme to its missing source and test it against 34 years of lottery data.

PNPriya Nair8 min readUpdated
Cover image for “Best Keno Numbers to Play: Myth vs Math”
On this page

The short answer

There are no best keno numbers to play. Every game draws 20 of the 80 numbers, so each number you pick has a 20/80 = 25% chance of hitting, every single draw. As Wizard of Odds (2024) puts it, "it makes no difference which numbers you choose." The famous lucky-number lists trace back to no data at all.

The meme with no methodology: 1, 4, 23, 34 and 72

Search for the best keno numbers to play and one list turns up everywhere: 1, 4, 23, 34 and 72. Its highest-profile carrier is PokerNews, whose 2024 keno number-strategy guide claims that "research suggests the most frequently drawn numbers in Keno are 1, 4, 23, 34, and 72."

Which research? The article never says. No dataset, no time window, no venue, no link. And a few paragraphs later, the same guide states that every number has an equal chance of being drawn because the game runs on a random number generator. The myth's biggest carrier can't sustain its own claim for the length of one page.

That's the pattern across the dozens of sites repeating the list: they copy each other, and none of them cites a source. It's one upstream rumor wearing many logos. The absence is the story here. Keno draws are public record, and official lotteries publish decades of frequency data, as you'll see below. If those five numbers actually ran hot anywhere, somebody would link the chart. Nobody does.

If the game itself is new to you, spots, catches, paytables, start with our overview of how keno works, and keep the keno terms glossary nearby for the vocabulary.

Are the best keno numbers the most drawn ones?

Over any long sample, no keno number is meaningfully more drawn than another; the "most drawn" number is just whichever one random noise nudged upward. The best public dataset we found, from Washington's Lottery (draws since November 1992, data through July 9, 2026), shows its all-time leader, 78, sitting only about 4% above average.

First, though, look at how the myth pages handle this question. The affiliate site Chipy runs a best-keno-numbers guide which, as indexed in search results (2026), names 27 the hottest number with a hit frequency of 25.59%. Hold that next to the base rate: every keno number appears in 25% of draws, because 20 of 80 come out each game. A "hottest number" at 25.59% isn't a tip. It's a confession that the numbers are uniform, dressed up as an edge. The same guide, as indexed, offers a separate hot list for land-based casinos that shares zero numbers with its online list.

Now the real data. Here are the extremes of Washington's Lottery's Daily Keno frequency chart, covering every draw from November 1992 through July 9, 2026:

Number

Hits since November 1992

Versus the ~3,074 expected

78 (most drawn)

3,196

about 4% above

23

3,173

about 3% above

38 (least drawn)

2,965

about 3.5% below

That window covers roughly 12,300 daily draws. At a 25% hit rate, every number should land near 3,074 hits, and a spread of minus 3.5% to plus 4% around that is exactly the size of wobble pure chance produces over a sample this big. Look at that number next to the base rate: it took 34 years of draws for the hottest number in the game to look just 4% hot.

One wrinkle worth being honest about: 23 appears in the meme list, and it does sit near the top of Washington's chart. But near the top after 34 years means a lifetime hit rate around 25.8% instead of 25%. If that's what a lucky number delivers, the label isn't buying you anything. Meanwhile the actual leader, 78, appears in nobody's lucky-number article. If frequency-chasing worked, every one of those pages would open with 78. None do.

Do hot and cold keno numbers matter?

No. Hot and cold numbers describe the recent past, and the next draw doesn't consult the past. Every game is an independent shuffle of all 80 numbers, so a number that's hit five games straight and a number that's missed twenty both carry the same 25% chance tonight.

This myth does arrive with official-looking paperwork, though. The Maryland Lottery runs a live keno hot-and-cold page, described as tracking the most and least frequently drawn numbers over the last 100 games. The Pennsylvania Lottery's keno page carries the same framing over the last 100 drawings (both live as of July 2026). When the lottery itself publishes the chart, playing the hot numbers stops feeling like superstition and starts feeling like homework.

Here's the catch: a 100-game window guarantees a hot list. Each number expects about 25 appearances in 100 games, and with 80 numbers on the board, a few will always drift up toward 30 and a few down toward 20 by luck alone. Somebody has to lead any leaderboard. These pages are engagement features, closer to a scoreboard than a forecast, and to the lotteries' credit, they don't claim predictive power.

Want to watch the illusion assemble and dissolve? The Connecticut Lottery's keno frequency tool (2026) lets you set the window yourself, from the last 10 drawings to the last 1,000. Short windows look streaky and dramatic. Long windows flatten everything toward the same 25%. The heat lives in the window you chose, not in the numbers.

Why do lucky numbers feel like they work?

Because betting on streaks is normal, well-documented human behavior, not a personal failing. The cleanest field evidence comes from Maryland itself. In 1993, economists Charles Clotfelter and Philip Cook published a study in Management Science using Maryland's daily-numbers betting records. Money bet on a number fell sharply right after that number was drawn, then took months to recover. Players treated a fresh winner as used up: the textbook gambler's fallacy, captured in real betting slips rather than a lab.

Three decades later, the same state's lottery publishes the keno hot-and-cold chart. The instinct didn't fade; it got a dashboard.

Memory does the rest of the work. You remember the night your birthday numbers swept the board and forget the hundreds of draws they sat out. Our brains are built to find patterns and keep selective score, which is also why rows, diagonals, and four-corner plays feel meaningful on the ticket grid. We take that folklore apart separately in keno winning patterns: myth vs math.

The math: every pick is a 25% shot

All of keno runs on one formula, the hypergeometric distribution: P(k) = C(20,k) x C(60,n-k) / C(80,n), the chance of catching k of the n numbers you picked. Set n = 1 and it collapses to 20/80, a flat 25% per number, per draw. Wild Academy computed the exact figures (2026), and no number on the board gets a different value.

Which numbers you choose changes nothing. How many you choose changes everything about the shape of your wins, which is why the keno odds and probability table is organized by spot count, not by number.

Could the machine itself play favorites? Ruling that out is what certification is for. Gaming Laboratories International, one of the industry's main testing labs, names the game directly in its RNG testing specifications: submissions must document their draw rules, "for example Keno, 20 numbers drawn from 1 through 80 without replacement," and the generator is tested for uniformity and bias. A certified keno game with genuinely hot numbers isn't a secret strategy. It's a failed audit.

So how should you pick your keno numbers?

However you want, because this is the one keno decision with no wrong answer. Birthdays, anniversaries, the quick-pick button, the same slip you've played for a decade: all carry identical expected value. Playing the same numbers every draw doesn't build up credit, and switching doesn't reset anything. Pick whatever makes checking your ticket fun.

Then spend your attention on the decisions that do move money. How many spots you play and which paytable you're facing set the house edge, the game's real price. Pace multiplies that price, the same bankroll arithmetic we walk through in the $20 slot method, where the staking story changes and the math never does. And if you're tempted to fix the numbers problem with a staking progression instead, keno betting systems runs the popular ones against keno's edge and reports the damage.

Numbers are free. Budgets aren't. Choose the first for fun and the second before you play.

Sources

Frequently asked questions

What are the best keno numbers to play?

There aren't any. Every keno draw takes 20 of the 80 numbers, so each number carries the same 20/80 = 25% chance per game, and Wizard of Odds (2024) is blunt: "it makes no difference which numbers you choose." Pick birthdays, use quick pick, or keep a favorite slip; the expected value is identical. What actually changes the math is how many spots you play and the paytable behind them.

What are the most drawn keno numbers?

Whichever ones random noise nudged upward in the window you happen to check. Washington's Lottery frequency data covering every Daily Keno draw from November 1992 through July 2026 shows the all-time leader, 78, at 3,196 hits versus 2,965 for last-place 38, roughly 4% above and 3.5% below the expected 3,074. That spread is ordinary chance over 12,300 draws, not a signal.

Do hot and cold numbers matter in keno?

No. Each draw is an independent shuffle of all 80 numbers, so recent streaks carry zero information about the next game. Lottery hot-and-cold pages, like the Maryland Lottery's 100-game keno tracker, are engagement features: over any 100 games each number expects about 25 hits, and a few will always drift high or low by luck alone. Connecticut Lottery's adjustable frequency tool shows the streaks flatten as the window grows.

Are 1, 4, 23, 34 and 72 really the luckiest keno numbers?

No. The list traces to no dataset. PokerNews (2024), its highest-profile carrier, says only that "research suggests" it, names no research, then states elsewhere on the same page that every number has an equal chance. Real long-run data disagrees too: the hottest number in Washington's Lottery's 34-year record is 78, which appears in none of these lucky-number lists.

Does playing the same numbers improve your chances?

No. Keno draws have no memory, so your regular numbers don't accumulate credit and never become "due." Clotfelter and Cook's 1993 Management Science study of Maryland lottery betting showed players do bet as if recent results matter, money on a just-drawn number dropped sharply, but the draw itself stays at 25% per number every game. Loyal numbers only make tickets quicker to check.

Go deeper with a course

About the author

Priya Nair

Table Games & Wallets Writer

Priya covers the classic tables — blackjack, roulette, baccarat, craps, and poker hand rankings — plus the wallet and self-custody basics every crypto player needs.

One clear lesson, every week

Join thousands learning crypto the right way. New guides and plain-English breakdowns — never spam, unsubscribe anytime.