Variance vs. RTP: Why Anything Can Happen in the Short Term

When a slot has an RTP of 96%, many players believe they will “on average get back 96 €” per 100 € wagered. That's true — but only over millions of spins. Over 100 spins, virtually any outcome from 0 € to well above 1,000 € is possible. The reason is variance.

RTP is the expected value: The expected value E(X) of a wager is the sum of every possible payout, weighted by its probability. For a 96%-RTP slot at a 1 € stake, E(X) = 0.96 €. You lose, in expectation, 0.04 € per spin — that's the house edge.

Variance is the swing around that value: Mathematically variance is σ² = E[(X − E(X))²] — the average squared distance from the expected value. The standard deviation σ is its square root. It tells you how far individual outcomes typically deviate from the mean.

Example: two games with identical RTP, completely different behaviour:

• Game A (low variance): roulette, even-money on red. RTP 97.3%, σ ≈ 1. You win or lose almost exactly 1 € per round — calm, predictable.

• Game B (high variance): a high-volatility slot with a 10,000× max win. RTP 96%, σ can exceed 50. You lose hundreds of spins in a row, then a 500× hit. Your session balance can swing by a factor of 100.

Confidence intervals: the real picture: The important question isn't “what's my expected profit?”, it's “how far can my actual result deviate from expectation?”. With n spins of single-spin variance σ², the standard deviation of the total result is √n × σ. Over 100 spins of a 96% slot with σ = 10, the typical swing is therefore ±100 € around an expected loss of −4 €. With 95% probability, your session result lands somewhere between −204 € and +196 €.

When does the RTP assert itself? Standard deviation grows with √n, expected loss grows with n. The relative ratio loss/swing therefore grows with √n. For a typical slot example, you need roughly 100,000 spins before the realised return is highly likely to land within ±1% of the theoretical RTP.

Practical consequences: First: “the slot is hot today” is not a property of the game, it's a short-term manifestation of variance. Tomorrow it can be “cold” without anything changing about the RTP. Second: high variance is neither good nor bad — it's a risk profile. Playing high-volatility slots on a small bankroll statistically means going bust much sooner. Third: betting systems like Martingale change neither expectation nor variance fundamentally — they just shift the distribution of outcomes, they don't make it favourable.

An honest rule of thumb: the higher the variance, the more luck feels like skill. That's exactly what makes high-volatility games attractive and at the same time more dangerous. Knowing the math lets you separate the two effects — and decide consciously what risk you actually want to take.