How slots actually work — and the real odds

Slots are the simplest game to play and the hardest to understand, because the one thing that decides whether you win — the paytable behind the reels — is hidden. This guide opens it up. Everything below is illustrated with our own transparent model; it shows how RTP and variance work, but it is not the odds of any real machine, which are proprietary and undisclosed.

What this guide covers

No strategy can change a slot's odds — there are no decisions after you press spin. So the useful knowledge is different from blackjack or poker: it's understanding what you're actually buying when you spin, so the machine can't fool you about it. By the end you'll be able to read any paytable, tell a "tight" machine from a "loose" one, and explain why a 96% machine can still empty your wallet in twenty minutes.

1. How a slot actually works

Forget the spinning reels for a moment — they're an animation. Underneath, every modern slot is just this:

  1. A random number generator picks one number, fairly, from a huge range.
  2. A paytable says which numbers win and how much. Most numbers win nothing; a few pay a little; a tiny sliver pays big.
  3. The reels then animate to show the result that was already decided. The "near miss" where the jackpot symbol lands just above the line is choreography, not chance.

Our model makes this literal: each spin draws one whole number from 1 to 1,000,000, and a paytable of number ranges maps it to a payout multiplier — ×0 (no win), ×1 (your stake back), or higher (a win). That's the entire machine. Real slots use the same RNG-and-paytable skeleton; they just hide the numbers and dress it in themes, reels, and bonus features.

2. RTP, hit rate & house edge

Three numbers describe any machine, and they're easy to mix up:

  • RTP (return to player) — the share of everything wagered that comes back as winnings over the long run. A 75.0% machine returns 75.0% of all money bet through it, eventually. The house edge is just the rest: 25.0%.
  • Hit rate — how often a spin returns at least your stake. This is not RTP: a machine can hit often with tiny wins and still have a low RTP, or hit rarely with huge wins and a high RTP.
  • Max win — the biggest multiplier on the machine. Big top prizes are paid for by colder, rarer wins everywhere else.

The key trap: RTP is a long-run average over millions of spins, not a promise about your session. Even our loosest machine (99.0% RTP) keeps 1.0% of every dollar forever — there is no setting at which a slot is a break-even bet. Watch the average converge in the Simulate tab: over thousands of spins the observed return creeps toward the exact RTP, but it's always below 100%.

3. Volatility: same RTP, different ride

This is the concept the casino never advertises. Volatility (or variance) is how swingy a machine is — and two machines with the identical RTP can feel like completely different games.

Our three 96% machines prove it. All three pay back exactly 96.0% over the long run:

  • Steady — frequent small wins, a smooth slow grind. Most sessions last a while.
  • Generous — long dry spells punctuated by the occasional big spike.
  • Wild — almost the entire 96.0% is hidden in one jackpot that hits about 1 in 8,333 spins (×5,000). Brutal droughts; the rare player who hits it walks away big, and almost everyone else busts.

Same house edge, wildly different experience. High volatility is why "this machine pays 96%" tells you almost nothing about what your night will look like. The Simulate tab's face-off runs all three on the same bankroll so you can see the bust rates side by side.

4. Read any machine's paytable

Here's every machine's full paytable, computed from its ranges — the exact thing a real casino keeps secret. Each row is an outcome: how often it happens, what it pays, and how much of the RTP it carries. Notice how much of the "return" on the high-volatility machines is locked inside outcomes you'll almost never see.

A real slot keeps its paytable secret. Ours is open: here is the exact chance of every outcome on every machine, computed at build time by the same solver the Play and Lab tabs use.

The lineup — what each machine keeps, and how wild the ride is

Machine RTP House edge Hit rate Net-win rate Max win Volatility
Tight machine (75% RTP) 75.0% 25.0% 32.9% 32.9% ×500 Medium (3.41)
Stingy machine (82% RTP) 82.0% 18.0% 12.6% 12.6% ×1,000 High (21.50)
Carnival machine (88% RTP) 88.0% 12.0% 42.0% 40.0% ×5 Low (1.12)
Standard machine (92% RTP) 92.0% 8.0% 35.5% 35.5% ×200 Medium (3.75)
Typical machine (94% RTP) 94.0% 6.0% 37.5% 35.5% ×150 Medium (3.41)
Steady 96% (low variance) 96.0% 4.0% 45.0% 45.0% ×8 Low (1.22)
Generous machine (96% RTP, high variance) 96.0% 4.0% 15.8% 15.8% ×2,000 High (30.41)
Wild 96% (extreme variance) 96.0% 4.0% 10.8% 10.8% ×5,000 Extreme (54.84)
Loose machine (99% RTP) 99.0% 1.0% 46.5% 46.5% ×6 Low (1.17)

The three machines highlighted in gold all pay back exactly 96% — identical long-run RTP, yet the volatility index runs from 1.22 (Steady) to 54.84 (Wild). Same house edge, completely different experience — that's the lesson the Lab makes you feel. Higher "max win" almost always means a colder, swingier ride.

Open the hood — the full paytable for one machine

About the middle of the pack for a modern machine.

RTP94.0%
House edge6.0%
Hit rate37.5%
Net-win rate35.5%
Max win×150
VolatilityMedium (3.41)

Every spin lands somewhere on that bar. The big block on the left is losing; the win bands are the thin strips on the right — proportional to how rare they actually are.

Outcome Numbers Chance Pays Share of return
No win 1–624,600 62.46% no win 0.0%
Stake back 624,601–644,600 2.00%1 in 50 ×1 (stake back) 2.0%
Small win 644,601–984,600 34.00% ×2 68.0%
Mid win 984,601–999,600 1.50%1 in 67 ×12 18.0%
Big win 999,601–1,000,000 0.040%1 in 2,500 ×150 6.0%
Total 1–1,000,000 100% 94.0%

"Share of return" is each band's contribution to the 94.0% RTP (chance × payout); the column sums to the RTP. Everything you don't get back is the 6.0% house edge.

Figures are exact for this illustrative model and shown rounded — not the odds of any real slot machine, which are proprietary and undisclosed. Percentages are rounded for display; the underlying ranges are exact integers over 1,000,000.

5. Bankroll reality

Because every spin is a small bet at a negative edge, the math is relentless: the more spins you buy, the closer your results track the house edge. "Time on device" is the industry's own metric — keep you spinning and the edge does the rest. A 25.0% edge sounds small until you realize a typical player cycles their bankroll many times an hour; each pass hands the house its cut again. The interactive house-edge model shows how even a 1% edge grinds a bankroll to zero given enough bets.

6. The myths that cost you money

  • "This machine is due." Spins are independent. A machine that hasn't paid in hours is exactly as likely to pay on the next spin as one that just hit. The RNG has no memory.
  • "Hot" and "cold" machines. Streaks are just variance — they're obvious looking backward and unpredictable looking forward.
  • Near misses mean you're close. The result was chosen before the reels moved; a jackpot symbol landing just off the line is animation designed to feel like a near miss. You weren't close.
  • Stopping the reels / bet patterns / lucky timing. None of it touches the drawn number. There is no input that changes a slot's odds.

7. Next steps

How this simulation works
Rules modeled
An illustrative slot: each spin draws one uniform integer in [1, 1,000,000]; a paytable of contiguous ranges maps it to a payout multiplier. Nine machines span 75%–99% RTP, each with a distinct volatility. Stakes $0.20–$5 in integer cents.
Assumptions
Spins are independent and identically distributed — no memory, no 'due' machine, no skill. NOT a model of any real slot's odds (those are proprietary and undisclosed); ours is fully open by design. Fun credits only, no real money.
Mathematical basis
RTP, hit frequency, net-win frequency, max win, and the volatility index are computed in closed form from the disclosed ranges and are exact for this model; a build-time test re-derives every machine's RTP and fails the deploy on drift. The Monte-Carlo Play/Lab validate against these exact values.
Engine version
slots engine 1.0
Validation
84 unit tests, incl. the build-time RTP verifier with a planted-broken fixture.
Last reviewed
Independently reviewed and PASSED (D-008). An independent reviewer re-derived every preset's RTP and volatility from the disclosed ranges, confirmed the outcome space tiles exactly, and verified the simulator's integer-cent accounting; the build-time integrity gate re-checks the figures on every deploy. Last reviewed 2026-06-29 (1 independent review pass).