How Many Bets Does It Take to Know Anything?

## The Core Problem Suppose you place 20 bets and show a profit. Are you skilled, or did you run well? The honest answer at 20 bets is: you cannot tell. This is not pessimism. It is mathematics. At small sample sizes, variance (luck) swamps skill. The signal is drowned in noise. ## What Variance Actually Means Variance is the natural spread of outcomes around the expected value. Even a coin weighted to land heads 55% of the time will land tails in long streaks. If you bet on it 20 times, tails can easily dominate by chance. Betting on football, where single-game outcomes are highly random, variance is enormous. Even a 10% edge on a market does not guarantee profit over 50 bets. It starts to show reliably around 500–1,000 bets in most markets. ## The Rough Numbers At 5% ROI, with average odds of 2.0: - At 50 bets: results are essentially noise - At 200 bets: a weak pattern starts to be detectable - At 500 bets: you can begin to have modest confidence in a genuine edge - At 1,000+ bets: statistical significance becomes achievable ## Implications for Evaluation Never evaluate a betting strategy on fewer than a few hundred bets. Never abandon a strategy during a normal variance downswing. Never conclude a strategy is working based on a short winning run. The uncomfortable truth is that most bettors never place enough bets in a consistent enough system to ever know whether they have an edge.