Betting Hit Rate and Confidence Intervals: What 60% Does Not Tell You
A 60-from-100 example uses a Wilson interval to show sampling uncertainty and why hit rate is not the same as profitability.
What a hit rate measures
Hit rate is the share of recorded bets that win: wins divided by the number of bets, multiplied by 100%. If 60 of 100 settled bets win, the point estimate is 60%. It describes that particular sample; it is not the exact future probability. With a small sample, a few extra wins or losses move the percentage substantially.
A confidence interval communicates uncertainty in the estimated proportion. The Wilson interval is useful for binomial outcomes. For 60 wins from 100, assuming independent and comparable observations, a standard two-sided 95% Wilson interval is approximately 50.2%–69.1%. This does not mean every future bet has a probability inside these bounds. It describes uncertainty around an overall proportion under the chosen sampling model. If odds, sports, markets, or settlement rules change, a single shared probability may not be a suitable assumption.
The Wilson formula uses p = w/n, where w is wins and n is the sample size, and z = 1.96 for the usual two-sided 95% level. Centre: (p + z²/(2n)) ÷ (1 + z²/n). Half-width: z × √(p(1−p)/n + z²/(4n²)) ÷ (1 + z²/n). The lower and upper bounds are the centre minus or plus the half-width. For p=0.60 and n=100, the result is about 0.502 and 0.691. A validated statistics library can do the arithmetic, but the method should be named rather than presented as a forecast.
Hit rate is not profit
A 60% win rate says nothing about profit without prices, stake sizes, pushes, and losses. At equal stakes, a hypothetical price of 1.60 can produce a different financial result from 2.20. Decide whether pushes are a separate category, and record all decisions before the event; missing selections and discarded losses can create bias. The guide to small betting samples explains why ten or twenty matches provide little evidence.
For a meaningful interval, define outcomes consistently, keep observations comparable, and publish inclusion rules in advance. Do not remove losses or pushes because they make the estimate look worse. Correlated bets or a sample selected after results may violate the assumptions behind a standard Wilson interval. Also see what a betting hit rate means. Neither a rate nor its interval guarantees future winners.
Practical check
Before calculating, define the sampling unit, treatment of pushes, and how half wins are counted. Three bets on one match are not necessarily three independent trials. Predefine the list of decisions: removing losing bets after results changes the estimate and makes an interval misleading. Report sample size, period, market, and win definition with the hit rate. An interval describes uncertainty in an estimate; it does not guarantee bounds for future bets.