Guide
Guide · 5 min read
Variance and sample size: why a hot streak proves little
A good run and a good bettor look identical for a surprisingly long time. The reason is variance: the natural swing of random results around their true average. Understanding variance is what lets you tell a real edge from a lucky streak, and it is the single most abused idea in a tout's sales pitch.
Why luck is loud at first
Flip a fair coin ten times and you will often see 7 heads or 3 heads, nowhere near a clean 5. That is not the coin being hot or cold. It is just how much random results bounce around over a few trials. Betting is the same. Over ten or twenty bets, a bettor with no edge at all can easily run well ahead or well behind, purely by chance. The small sample is loud, and it says almost nothing about skill.
How the swing shrinks as bets add up
Over 10 coin-flip bets, finishing 7 and 3 is common and means nothing.
Over 100 bets, results cluster closer to the true rate, but a few points of swing remain.
Over 1,000 bets, the average is much steadier, and a real edge starts to show through.
The more bets, the quieter the luck, which is why big samples are the only honest test.
How many bets it takes
Because the edges in betting are small, it takes hundreds of bets, often more than a thousand, before results separate skill from luck with any confidence. That is far more than a single season of a tout's plays. It is why a 20-and-5 run posted as proof is not proof at all, and why the honest measures of skill are the ones that do not need a big winning sample to read, like the price you took against the closing number.
What a large study showed
Why price beats record over a big sample
164,283 real bets, large enough to quiet most of the luck.
Bets whose price beat the closing number returned +$3.70 per $100.
Bets whose price did not returned −$5.34 per $100.
The split held because the sample was big. On a handful of bets, it would not.
Sorted into grades against the closing number, the same big sample separates cleanly: the best-priced group returned +$5.63 per $100, while the worst returned −$7.13. That separation only shows up because there are enough bets to drown out the noise. Over a short streak, the grades would look like random scatter, which is exactly why a streak proves so little.
How to stay honest with yourself
Variance cuts both ways, and it works on your own feelings as much as on a tout's ad. A cold streak does not mean your approach is broken, and a hot one does not mean it is working. Keep enough records that you can look at a few hundred bets at once, judge yourself on the prices you took rather than the nights you happened to win, and treat any short run, good or bad, as mostly noise until the sample grows. That patience is what separates a measured bettor from a chasing one.
Variance is the swing of random results, and it is loud over a few bets and quiet over many. It takes hundreds of bets to see skill, so a hot streak proves little, and the price you took is a cleaner signal than any short record.
For entertainment and education. Nothing here predicts results or recommends a bet.