The House Always Wins: The Math Behind the Certainty of Loss
Why expected value and the law of large numbers guarantee that every casino game is a losing proposition over time — and why no strategy changes this.
Every casino game — whether it runs on a blockchain or a carpeted floor in Las Vegas — shares one inescapable mathematical property: the house expects to keep a percentage of every bet, forever. This isn’t a conspiracy or a design flaw. It is the intended, engineered outcome, and understanding the arithmetic behind it is the single most important thing anyone can know before wagering anything of value.
What Is the House Edge?
The house edge is the percentage of each bet the casino expects to retain as profit, calculated over an infinite number of rounds. It exists because the payout odds are set slightly below the true probability of winning.
A simple example: imagine a coin-flip game where heads pays 1.8× your stake and tails loses your stake. The true probability of heads is 50%, so a fair payout would be 2×. The gap between 2× and 1.8× is the house edge — in this case 10%. For every £100 wagered, the casino expects to keep £10.
Common house edges in crypto casinos:
| Game | Typical House Edge |
|---|---|
| Roulette (European) | 2.7% |
| Roulette (American) | 5.26% |
| Blackjack (optimal play) | 0.5–1% |
| Slots | 3–15% |
| Dice (e.g., 2% edge) | 2% |
| Crash (varies by multiplier) | 1–4% |
Provably fair games sometimes advertise lower edges, but the edge is always present. See our provably fair explainer for how to verify the stated RTP yourself.
Expected Value: The Engine of Guaranteed Loss
Expected value (EV) is the average outcome of a bet if it were repeated an infinite number of times. A negative expected value means you are mathematically expected to lose money. Every casino game has a negative EV for the player.
The formula is straightforward:
EV = (Probability of winning × Profit) − (Probability of losing × Stake)
Using our coin-flip example:
- EV = (0.5 × 0.8) − (0.5 × 1) = 0.4 − 0.5 = −£0.10 per £1 wagered
This is not a bad run of luck. This is the mathematical certainty of what will happen if you play long enough.
The Law of Large Numbers: Why Time Is the Casino’s Best Friend
The law of large numbers states that as the number of trials increases, the observed outcome converges toward the expected outcome. In the short run, gamblers win and lose unpredictably. But as sessions accumulate, the casino’s take converges reliably toward its stated edge.
This is why casinos have no reason to cheat. The math works for them automatically. A player who wagers £100,000 in total across their lifetime on a game with a 3% house edge will, on average, lose £3,000 of it. The only variable is timing.
High-frequency crypto gambling accelerates this process considerably. A player placing a new bet every five seconds on a dice site with a 2% edge, wagering £1 per bet, expects to lose roughly £1,440 per hour in pure EV terms — before variance even enters the picture.
The Gambler’s Fallacy
The gambler’s fallacy is the mistaken belief that past outcomes influence future independent events. After ten consecutive reds on roulette, many players bet heavily on black, feeling it is “due.” It is not.
Each spin is statistically independent. The roulette wheel has no memory. The probability of red or black on the next spin is identical whether it followed ten reds or ten blacks. Believing otherwise is not just an error — it is the belief pattern that sustains loss-chasing behaviour and keeps players at the table far longer than they intended.
Crypto games make this worse by providing visible history panels, streak displays, and real-time chat where other players reinforce fallacious reasoning (“it hasn’t crashed above 10× in 200 rounds, it’s overdue”).
Variance Is Not Your Friend Either
Variance is the statistical spread around the expected value. High-variance games (slots, crash, high-multiplier dice) produce dramatic wins and losses in the short run while still converging toward the same negative expected outcome over time.
High variance is appealing because it occasionally delivers large wins. But those wins do not change the underlying economics — they borrow from future sessions. Every win is, in the long run, offset by proportionally larger accumulated losses. The bigger the jackpot, the worse the return-to-player percentage tends to be.
What This Means in Practice
- No session length makes gambling profitable. Quitting while ahead is the only way to bank a win — and the desire to keep playing makes that psychologically difficult by design.
- Bonuses do not fix negative EV. They add conditions (wagering requirements, game restrictions) that often make the true EV even more negative. See our article on bonus traps for more detail.
- “Provably fair” proves randomness, not fairness. A provably fair game still has a house edge; it merely proves the randomness of outcomes cannot be manipulated. Read the provably fair guide to understand what the verification actually covers.
Understanding the mathematics of expected value is not intended to be discouraging — it is intended to be clarifying. If you choose to gamble, do so as a form of entertainment with money you have decided to spend, not as an investment or a strategy. For support with gambling behaviour, visit our responsible gambling page.