Analytics and intelligence, bias modeling for the 3rd team on the field.
Ranked NFL officiating patterns from the Ref Watch dataset, with effect sizes, sample sizes, and links to evidence.
8 high-confidence findings
League baseline
46.0 combined points12.5 flags thrown per gamebenchmark 44.2 (O/U comparison threshold)
#1
League trend
Sample: 571 games • Data maturity: High
46.0 points2025-26 avg total
286 games
Year-over-year context from 10 seasons (2,761 total games). League-wide baselines, not ref-specific. Sample depth is strong (100+ games in this finding).
#2
Marquee efficiency
63.3%Marquee over rate
79 marquee games · vs 14.3% baseline
Sample: 79 games • Data maturity: High
Prime-time and high-profile slate: 63.3% over rate (79 games) vs 14.3% baseline, +49.0 pts.
View evidence#3
Whistle extreme
17.2Flags per game
+4.9 vs 12.4 league avg
Sample: 45 games • Data maturity: High
Alex Moore averages 17.2 flags per game (+4.9 vs the 12.4 league average) across 45 assigned games, one of the clearest penalty-volume outliers in the NFL pool.
View evidence#4
ATS historical tendency
30.0%ATS cover
9-19-2
Sample: 30 games • Data maturity: High
Across 30 lined games, teams are 30.0% against the spread with James Carter II - 20.0 pts from a neutral 50% split, independent of straight-up wins.
View evidence#5
O/U historical tendency
24-7O/U ATS
77.4%
Sample: 31 games • Data maturity: High
Totals go overs 77.4% (24-7) across 31 lined games.
View evidence#6
Whistle extreme
15flags per game
+2.6 vs 12.4 league avg
Sample: 52 games • Data maturity: High
Byron Boston averages 15 flags per game (+2.6 vs league) across 52 games, the largest whistle delta among 50+ game refs.
View evidence#7
Ref outlier
+3.8Scoring delta
vs 45.6 league avg
Sample: 87 games • Data maturity: High
Mark Hittner's 87 games average 49.3 combined points (63.2% over 45.6), one of the largest scoring deltas in the pool.
View evidence#8
Ref outlier
44.8%Under benchmark
87 games
Sample: 87 games • Data maturity: High
44.8% of Jimmy Russell's 87 games finish under 45.6 combined total points, 5.2 pts from a neutral 50% baseline.
View evidenceThese are descriptive historical associations. They do not establish causation or predict individual games.