Slate Stats › NFL › Studies
Does the referee affect fantasy scoring?
Verdict — Predictable, and far too smallBarely. A referee’s flag rate is one of the most reproducible things in football — more predictable than any defensive rating we have measured — and the busiest crews are worth about one extra pass play per team per game. For a team’s top receiver that is roughly a sixth of a catch.
This is the best example we have of a trap that catches analysts constantly: a signal can be extremely consistent and still be far too small to act on. Referee tendencies are both at once, which makes them a useful thing to measure once and then stop thinking about.
| Referee third | Flags per game | Plays | Pass plays | Per team |
|---|---|---|---|---|
| Quietest | 11.7 | 174.1 | 72.4 | 36.2 |
| Busiest | 13.3 | 177.0 | 74.4 | 37.2 |
| Difference | +1.6 | +2.9 | +2.0 | +1.0 |
Across 27 referees with at least 20 games from 2016 to 2025, flag rates run from 10.8 to 15.0 per game — a wide, stable spread. Split the same referee’s games into odd and even weeks and the two halves agree at +0.80, which corrects to +0.89. That is higher than anything a defense does.
Then it collapses on contact with the thing you would actually bet or start a player on. Games under the busiest third of referees run 177.0 plays against 174.1 under the quietest third, and 74.4 pass plays against 72.4. That is 2.0 extra pass plays per game, or about one per team.
One extra pass play is roughly one extra attempt. Spread across a team’s receivers by their usual target share, the number one option collects about a quarter of a target, which at a normal catch rate is a sixth of a reception. No lineup decision survives contact with a number that size.
The lesson generalises and it is the reason this page exists. Reliability tells you a quantity is measurable. It does not tell you the quantity is large. Several of the most reproducible things in football — referee habits, how often a defense plays man coverage, which route types it concedes — turn out to move a receiver by about a tenth of a catch. Measure reliability first, then always ask how big.
How this was measured
Every regular-season game from 2016 through 2025 whose crew we could identify, joined to play-by-play. The officials feed keys on the legacy numeric game id and reaches the rest of the data only through a translation column, which is why it had gone unused. 2,758 referee-games matched. Reliability is a split-half correlation of flag rate across odd and even weeks for referees with at least 10 games in each half (n=27), corrected by Spearman-Brown. Volume figures compare the top and bottom third of referees by flag rate.