Stake Keno Guide: Pick Counts, Risk Levels, Payout Tables

Published 2026-04-16 · Tags: guides, keno, strategy

Stake's Keno is the lowest-RTP Original (92-95%). Full breakdown of each risk level, real hit probabilities, and the most honest Keno config to play.

Keno on Stake looks like a casual click-and-wait game. In reality it has the lowest RTP of the Originals (varies by risk level, bottom out at 92%) and its payout tables are the most opaque. This post decodes every config, shows which pick counts are reasonable, and explains why "Classic" mode is the honest default.

How Stake Keno works

You're looking at a 40-number grid. Process:

1. Pick 1 to 10 numbers you're betting on
2. Set your stake
3. Set your risk level: Classic, Low, Medium, or High
4. The game randomly draws 10 numbers from the 40 using HMAC-SHA256 with removal (once drawn, a number can't be drawn again)
5. Your payout depends on how many of your picks were in the drawn 10, cross-referenced with your risk level's table

It's fully provably fair, you can verify any round with our Stake Verifier, which supports all four Keno risk levels.

The actual hit probabilities

The probability of hitting X of your Y picks when 10 balls are drawn from 40 follows a hypergeometric distribution. Here are the exact percentages for common pick counts:

5 picks (out of 40), 10 drawn

• 0 hits: 20.5%
• 1 hit: 37.3%
• 2 hits: 28.6%
• 3 hits: 10.8%
• 4 hits: 2.4%
• 5 hits: 0.25% (jackpot)

8 picks

• 0 hits: 4.7%
• 1 hit: 17.0%
• 2 hits: 26.5%
• 3 hits: 23.1%
• 4 hits: 13.2%
• 5 hits: 4.9%
• 6 hits: 1.1%
• 7 hits: 0.14%
• 8 hits: 0.0075% (jackpot, basically 1 in 13,000)

10 picks

• 0 hits: 0.38%
• 1 hit: 2.3%
• 2 hits: 7.0%
• 3 hits: 13.9%
• 4 hits: 19.0%
• 5 hits: 18.8%
• 6 hits: 13.5%
• 7 hits: 6.8%
• 8 hits: 2.3%
• 9 hits: 0.47%
• 10 hits: 0.036% (jackpot, ~1 in 2,800, advertised big on streams; almost never happens)

The four risk levels

Classic

Closest to an honest keno experience. Most hits pay something (small), and the jackpot pays meaningfully but not a thousand times your stake. Narrow variance, long-lasting bankroll. RTP: ~94% for most pick counts.

Low

Smoother payouts, emphasizing partial-hit wins. 0-hit bets often return 0.5–1× stake. Very tight variance. RTP: ~92%–94%.

Medium

Jackpot multipliers expand, but 0-hit outcomes pay 0 (straight loss). More volatile. RTP: ~94%–95%.

High

Huge jackpot multipliers (thousands of times stake for max hits), but small-hit outcomes pay poorly or not at all. Massive variance. Jackpots are so rare that most sessions are pure losses. RTP: ~93%–94%.

So Keno is one of Stake's worst-RTP Originals

Yes. Keno is 92–95% RTP depending on config, compared to 99% for Dice, Limbo, and Plinko. That's a 5–7% bigger house edge.

In practice, a $1,000-volume Keno session loses about $60 in expectation, vs $10 on Dice. At scale this adds up fast.

If you wonder why Keno feels rougher than other Originals: because it is. The RTP is published per config in-game (info icon → "Provably Fair") — check it.

Why people still play it

• Big-multiplier fantasy. 10-pick High risk can pay 10,000× stake on a full 10-hit. It never happens (1 in ~2,800 chance) but the theoretical ceiling is intoxicating.
• Slow, deliberate pace. No auto-spin tilt. You pick, draw, get result, deliberate clicks. Some players find this relaxing.
• Grid aesthetics. The visual design is satisfying even when losing.

These are entertainment reasons, not math reasons. Accept the worse RTP if the vibe's what you want.

The pick-count math

Does picking 2 numbers vs 10 numbers change your EV? No, all pick counts at the same risk level have the same RTP. The payout tables are tuned so the expected return matches regardless of how many numbers you pick.

What changes is variance:

• 2 picks: narrow variance, small wins/losses
• 10 picks: huge variance, massive possible jackpot, most rounds are 0-return

A reasonable middle ground: 5–7 picks at Classic risk. You get the "something hit" feeling most rounds without lottery-chasing.

The "pick adjacent numbers" myth

Some players believe numbers cluster — picking 3, 4, 5, 6, 7 is "better" than 2, 11, 23, 35, 39.

Not true. Every number has equal probability of being drawn. Clustering your picks has no effect. The only thing your pick selection affects is how much variance you see in your own session, because you're essentially picking a subset of the 40-ball universe.

(Minor note: the distribution of matches differs if all your picks are adjacent vs spread, due to the hypergeometric nature. But the mean is identical. Pick whatever makes you happy; it won't change your long-run RTP.)

The auto-pick question

Stake's auto-pick button picks 5–10 random numbers for you. Functionally equivalent to you picking them manually, same EV, same RTP. Auto-pick is for players who don't want to think about which numbers; there's no mathematical advantage either way.

Reasonable Keno playbook

If you're going to play Keno:

Pick count: 5–7 gives the best balance between engagement and bankroll survival. Avoid 10-picks unless you want pure lottery-chasing.

Risk level: Classic is the honest default. Low is for the smoothest experience; High only if you're chasing jackpots and accept most rounds will pay $0.

Session size: Cap your round count (say, 100–200 rounds), because each round is a fresh opportunity for small losses. Don't let time blur your round count.

Stake sizing: Stay under 1% of bankroll per round. Keno's variance at High risk can surprise you.

Stop-loss: Set it tight (25–50% of session budget). If Keno's not paying today, walking is the right move; the distribution doesn't care whether you're "overdue."

Verify your Keno results

If a session felt unusual:

1. Rotate your server seed after the session to reveal the old one
2. Open our Stake Verifier, pick "Keno" and the matching risk level
3. Paste the revealed seed + your client seed + the nonce range
4. For each nonce the verifier computes the 10 drawn numbers and your payout

Every round should match exactly. If any mismatch → you've caught a real issue (this has never happened on Stake in practice, but verification is free).

The verifier also has a Deep Dive feature for Keno — it'll scan up to 10 million nonces for high-hit outcomes (e.g. "find every nonce in my range where 8 or more of my numbers hit"). Good for after-action analysis of "did I almost hit the jackpot?" thoughts.

Our data on Keno

We track thousands of Stake users via the Bet Tracker. Keno-heavy players consistently underperform Dice/Limbo players by the RTP gap, about 5–7% worse P/L per unit volume on average. The game's popularity is driven by vibe, not by statistical performance.

The point

Keno is Stake's lowest-RTP Original. It's fully fair, just structurally worse at returning your stake than Dice, Limbo, or Mines. Play it for the aesthetics and the jackpot fantasy; don't confuse the feel with the expected value. Classic risk + 5–7 picks is the most honest config. Pick small, cap rounds, walk.

Scan any Keno nonce range for jackpot hits. Open the Stake Verifier →