How many bets until you know whether you are any good
The uncomfortable answer is more than most people ever place. What sample size is needed to distinguish a real edge from variance, why longer odds need far larger samples, and the one measurement that gives an answer in a fraction of the time.
At average odds of 2.00, distinguishing a genuine 3% edge from break-even with reasonable confidence takes roughly 1,000 settled bets. At odds of 5.00 it takes several times that. Most bettors never reach the sample size at which their profit figure means anything — which is why closing line value, which converges much faster, is the measurement professionals actually track.
Why does it take so many bets?
Betting outcomes are binary and the payoffs are uneven, so the variance of the return per bet is large relative to the edge being measured. The standard error of measured ROI falls only with the square root of the sample: quadrupling the number of bets halves the uncertainty.
| Average odds | Bets needed to detect a 3% edge | Bets needed for a 5% edge |
|---|---|---|
| 1.50 | ~450 | ~160 |
| 2.00 | ~1,100 | ~400 |
| 3.00 | ~2,200 | ~800 |
| 5.00 | ~4,400 | ~1,600 |
| 10.00 | ~10,000 | ~3,600 |
These are orders of magnitude, not precise thresholds — they assume flat stakes and independent bets. The pattern is what matters: the longer the prices you take, the longer you must wait to learn anything about yourself.
What does a losing run mean, and what does it not?
Runs that feel like evidence usually are not. For independent bets at a given strike rate, a run of k consecutive losses somewhere inside n bets becomes likely surprisingly quickly:
| Strike rate | Bets | Losing run that is unremarkable |
|---|---|---|
| 60% | 100 | 5–6 in a row |
| 60% | 500 | 7–8 in a row |
| 50% | 100 | 6–7 in a row |
| 50% | 500 | 9–10 in a row |
| 40% | 500 | 12–13 in a row |
Read the table before the run happens rather than during it. A sequence that looks like proof a method has stopped working is, at these frequencies, an ordinary event.
The faster measurement
Closing line value asks a different question: not "did this bet win" but "was the price better than the market's final price". Because it produces a signal on every bet regardless of outcome, it converges in a few hundred bets rather than a few thousand.
It is not a complete substitute. Beating the closing line consistently is strong evidence of skill; failing to beat it while still profiting is possible over short samples and usually is variance. But as an early read on whether a process is working, nothing else comes close — see closing line value.
The same limits apply to any published record, ours included. This is why our front page carries sample size next to profit and drawdown rather than a headline percentage: at small samples the percentage is the least informative of the three.
Check any of this against our record
Every signal CONSENSUS publishes carries the bookmaker odds fixed before the event starts and the settled result afterwards — including the drawdowns and the losing runs. The running total is on the front page and every entry is in the log.