The Math Behind Why You Can't Beat a Casino

Written by Beat the Spin Editorial · Published 2026-04-23

If you play $100 a night of slots with a 4% house edge, for a year, the math says you will lose about $1,460. This isn't a prediction that might go wrong — it's an expected value that, over enough sessions, you converge to with near-certainty. Here's why the number is unavoidable, and why no betting system changes it.

House edge: the cost of playing

The house edge is the average fraction of every dollar wagered that the casino keeps. On an American roulette wheel with 38 pockets (1-36, plus 0 and 00), a $1 bet on red pays $1 if you win and costs $1 if you lose. Red hits 18 of 38 times. Your expected return on that bet is:

(18/38) × $1 − (20/38) × $1 = −$0.0526

For every $1 wagered, the house takes 5.26¢ on average. That 5.26% is the house edge of American roulette. It doesn't fluctuate with streaks, betting patterns, or "hot" tables. It's baked into the wheel.

Edge values are typical. Individual games and pay-table variants differ; ranges span 2-10% for slots and 0.3-1.5% for blackjack depending on rules.

RTP, hit frequency, and volatility are three different things

Slot marketing material often cites "RTP" (return to player) as if it were the whole picture. It isn't. RTP is the long-run average return — a 96% RTP slot gives back $96 per $100 wagered over millions of spins. But two slots with identical RTP can feel radically different over one session because they pay out differently. Three concepts matter:

RTP — long-run average return. Inverse of house edge. RTP 96% = edge 4%.
Hit frequency — fraction of spins that return anything. A slot can have 30% hit frequency (you win something on 3 of every 10 spins) or 10% (you win on 1 of every 10).
Volatility — how spread out the wins are. Low-volatility slots pay small amounts often; high-volatility slots go dry for hundreds of spins then pay a jackpot.

Two slots at the same RTP but different volatilities feel nothing alike:

Slot	RTP	Hit freq	Volatility	Feel
Low-vol	96%	30%	Low	Frequent small wins, gradual bleed
Med-vol	96%	18%	Medium	Occasional decent win, regular dry spells
High-vol	96%	8%	High	Long dry spells, rare large payouts

All three lose the same amount long-term. The difference is whether you notice it spin-by-spin or discover it at the end of the night.

Expected value: the formula

Expected value (EV) is the average outcome of a bet, weighted by probability. For a simple bet with one win outcome and one loss outcome:

EV = (P_win × payout) − (P_loss × bet)

Consider a blackjack hand played with basic strategy at typical house rules. The house edge is about 0.5%, so on a $10 bet the EV is:

EV = $10 × (−0.005) = −$0.05

Every $10 hand costs you 5 cents on average. That is the smallest casino edge of any major game, and it still means a player betting $10 a hand for 100 hands walks away, on average, $5 lighter.

Why you can't beat it long-term

The key math result is the law of large numbers: as the number of bets grows, the average outcome converges to the expected value. The variance that makes a single session feel lucky or unlucky shrinks relative to the total amount wagered. Over 100 spins, anything can happen. Over 10,000 spins, you're nearly guaranteed to be within a percent or two of the EV line.

Illustrative. Real sessions vary; shorter sessions have wider bands around the EV line.

Want to see this math applied to a specific bonus?

Run our EV calculator → — enter any bonus amount, playthrough, and game type, see the expected outcome distribution over 10,000 simulated sessions.

Common misconception: the gambler's fallacy

"Red has come up seven times in a row; black is due." This is the gambler's fallacy — the belief that past outcomes change the probability of future outcomes in an independent random process. They don't. The roulette wheel has no memory. Each spin has the same 18/38 chance of red. Seven reds in a row doesn't shift the probability of the next spin; the wheel is no more "due" for black than it was before.

The same fallacy underlies progressive betting systems (Martingale, Labouchere, d'Alembert). None of them change the EV of the underlying bet. Doubling your bet after a loss doesn't flip the math — it just makes one ruinous losing streak catastrophic, which at any casino table is inevitable given enough time.

When the math doesn't apply

The inevitability of house-edge losses applies to games where the casino is the house. It doesn't apply, or applies differently, to:

Poker against other players — the casino takes a rake, but your edge vs. other players is skill-dependent. Winning players exist long-term.
Sports betting with a positive-EV edge — if you can identify lines where the bookmaker has mispriced the probability, you can beat the vig. This is hard and competitive; most bettors can't.
Advantage play — card counting, video poker full-pay machines, promotional arbitrage. Edges exist but are narrow and actively countered by casinos.
Sweepstakes-model social casinos — the math still applies to the game outcomes, but the bonus structure (free SC via daily logins, mail-in entries) means you can play at negative cost if you farm bonuses efficiently. See our state-by-state legal guide before deciding whether to participate.