How Long Until You Know?
Suppose you really do have an edge. How long before your results prove it? Probably much longer than you think.
Here, edge means expected ROI. A true +3% edge means every $100 bet is worth about +$3 on average before variance.
True win probability at +3.0% · 53.95%Make a guess.
Every bettor below really has a +3.0% edge.
After 100 bets, how many of the 100 do you think will still be losing money?
How long until you actually know?
At +3.0%, approximatelybefore roughly 95 out of 100 bettors with this edge would be expected to be above break-even.
This is an approximation from a normal model, not an exact sample-size law.
Same edge. More variance.
Longer prices need larger samples because the wins arrive less often and pay more when they do.
A hundred bets feels big. Statistically, it isn’t.
If you are down after 100 bets, that is not strong evidence that a small real edge is absent.
If you are up after 100 bets, that is not strong evidence that the edge is real either.
The smaller the edge, the more easily short-term luck can hide it.
Every bettor in this simulation has the same genuine edge. A bad six-week stretch can happen even when an edge is real — and a good stretch is not proof that an edge exists.
Results become useful slowly. Price information arrives much faster.
That is why I track the price I took against the closing market instead of waiting for win rate to tell me everything.
Play Beat The Close →Every model in one bundle.
See all models