Analytics and intelligence, bias modeling for the 3rd team on the field.
Ranked NCAA Football officiating patterns from the Ref Watch dataset, with effect sizes, sample sizes, and links to evidence.
6 high-confidence findings
League baseline
51.5 combined points11.0 flags thrown per gamebenchmark 50.6 (O/U comparison threshold)
#1
League trend
Sample: 1,426 games • Data maturity: High
1,426Total games
6 seasons
Game-level scores and penalties are live. Ref-level findings populate when ESPN crew assignments are available. Sample depth is strong (100+ games in this finding).
#2
Scoring extreme
51.5Coldest season
2025-26
Sample: 419 games • Data maturity: High
51.5 combined points per game vs 60.4 in 2020-21, a 8.9-point spread across seasons.
View evidence#3
League trend
60.42020-21 avg
165 games
Sample: 419 games • Data maturity: High
League average fell from 60.4 points in 2020-21 to 51.5 in 2025-26 across live-conference games.
View evidence#4
Scoring extreme
60.4Peak season avg
2020-21
Sample: 165 games • Data maturity: High
60.4 combined points per game across 165 live-conference games, 8.9 above the latest season.
View evidence#5
League trend
12.22020-21 flags
Per game
Sample: 419 games • Data maturity: High
Flags dropped from 12.2 per game (2020-21) to 11.0 (2025-26), a -1.2 change.
View evidence#6
League trend
51.5 points2025-26 avg total
254 games
Sample: 508 games • Data maturity: High
League scoring decreased by 0.4 pts year over year (2024-25 to 2025-26).
View evidenceThese are descriptive historical associations. They do not establish causation or predict individual games.