Guide
Why sample size matters in betting statistics
Almost every misleading betting statistic online is a true number reported without its sample size. A 40% return on investment from twelve bets is noise wearing the costume of evidence.
How much variance to expect
At even-money prices, a strategy with a genuine 3% edge still has a meaningful chance of showing a loss after 100 bets. The signal is small relative to the noise; only volume separates them.
As a rough working rule: under 100 settled bets, treat a record as an illustration of the process rather than proof of an edge. Several hundred settled bets are needed before return on investment becomes a serious estimate.
How to read a performance table
Four checks catch most of the misleading records you will encounter.
- Look at the settled-bet count before the profit figure.
- Check that the same selections were tracked across every staking plan being compared.
- Look for the drawdown, not just the final bankroll.
- Confirm results are recorded at bet time, not selected after the fact.
How this site handles it
Every published statistic comes from bets recorded by the robot before kick-off and settled from public final scores. The settled-bet count is displayed next to every figure, and no strategy is described as profitable while the sample is too small to support that claim.