Player stats are seductive: clean numbers, easy to compare, easy to argue from. But most player stats are noisier than they look, and the noise is concentrated in the moments that matter most for fantasy picks.
Sample size is everything
A player with 200 career IPL innings and a 32 average is statistically a different bet than a player with 30 innings and a 38 average. The smaller sample has more variance — the 38 average could be a hot streak that will revert, or a genuine step up. Most player-evaluation mistakes come from overweighting small samples.
Use this rule: any stat based on fewer than 50 innings or fewer than 100 overs bowled is a hint, not a conclusion. Wait for the sample to grow before betting on it.
Recent form is not a stat
The last six innings are not a statistically reliable predictor of the next six. The variance on six-innings samples is enormous. Recent-form-based picks are gambling on a small-sample trend, not on a stable skill. Some nights the trend continues; many nights it does not.
Use this rule: if a player’s last six innings look very different from their last 30, ask why. If the reason is a genuine change (injury recovery, role change, new batting position), the recent form is informative. If the reason is just hot or cold variance, treat it as noise.
Venue and matchup splits
Venue and matchup splits are useful but tricky. A player with a great record at the Wankhede may have played only 5 innings there; that is not enough to call it a strength. A player with a great record against Rashid Khan may have played against him only twice; that is not a stable edge.
Use this rule: any venue or matchup split based on fewer than 10 observations is a hint, not a conclusion. If you cannot find 10 observations, the split is not large enough to be meaningful.