The Pythagorean Baseline: Understanding Tempo-Free Efficiency
The Pythagorean Baseline
You may have noticed our Pythagorean rankings looking strangeāteams with high-scoring offenses ranking way too high, regardless of their actual quality. The original model was using raw points to estimate team strength. Today, we've deployed a proper efficiency-based Pythagorean model that aligns with more opponent-aware efficiency systems.
What Was Wrong
The original formula was simple:
Win% = Points^14 / (Points^14 + Points_Allowed^14)
This has a critical flaw: pace bias. A fast-tempo team that scores 90 and allows 85 looks the same as a slow-tempo team that scores 60 and allows 55. But the first team might be playing a fundamentally different gameāmore possessions, more variance, and not necessarily better.
The Fix: Efficiency-Based Pythagorean
We now calculate points per possession (or more precisely, points per 100 possessions) for both offense and defense:
# Estimate possessions
possessions = FGA - ORB + TOV + 0.475 * FTA
# Calculate efficiency
Offensive_Efficiency = (Points / Possessions) * 100
Defensive_Efficiency = (Opp_Points / Possessions) * 100
# Pythagorean formula with efficiency
Win% = OE^11.5 / (OE^11.5 + DE^11.5)
Key Details
-
Possession Estimation: We use the standard formula from basketball analytics. The 0.475 multiplier on free throw attempts accounts for and-one situations and technical fouls.
-
Offensive Rebounding: Since we only have total rebounds in our data, we estimate ORB as ~27% of total rebounds (league average).
-
Exponent: We use 11.5 for college basketball rather than the 14.0 typically used for NBA.
The New Rankings (Jan 2026)
| Rank | Team | OE | DE | Net | Win% |
|---|---|---|---|---|---|
| 1 | Iowa State Cyclones | 127.0 | 87.4 | +39.6 | 98.7% |
| 2 | Michigan Wolverines | 120.2 | 84.7 | +35.5 | 98.2% |
| 3 | Iowa Hawkeyes | 126.3 | 90.2 | +36.0 | 97.9% |
| 4 | Saint Louis Billikens | 124.1 | 89.3 | +34.8 | 97.8% |
| 5 | Duke Blue Devils | 121.1 | 87.6 | +33.5 | 97.7% |
| 6 | Tennessee Volunteers | 119.0 | 86.4 | +32.6 | 97.5% |
| 7 | Louisville Cardinals | 122.5 | 90.0 | +32.5 | 97.2% |
| 8 | Arizona Wildcats | 119.9 | 88.3 | +31.6 | 97.1% |
| 9 | Purdue Boilermakers | 126.4 | 93.1 | +33.2 | 97.1% |
| 10 | Georgia Bulldogs | 119.1 | 88.3 | +30.8 | 96.9% |
What This Baseline Tells You
The Pythagorean model is intentionally simple. It does not: - Adjust for opponent strength (like adjusted efficiency does) - Account for home court advantage - Weight recent games more heavily
This is by design. It's a transparent baseline that shows what a team's efficiency looks like in raw form. If you see a mid-major with elite Pythagorean numbers, they might be beating up on weak competitionāor they might be genuinely underrated.
Refresh Cadence
The model now runs nightly at 2:30 AM PST, after our ELO model, and recalculates season-level efficiency ratings.
You can see the current Pythagorean rankings on the main Rankings page, alongside our ELO ratings.
Next Steps
This baseline opens the door for more sophisticated models: - Adjusted Efficiency: Weight each game by opponent strength - Four Factors Analysis: Break down efficiency into shooting, turnovers, rebounding, and free throws - Tempo Analysis: Separate pace from efficiency to understand playing style
For now, the Pythagorean baseline gives us a clean, interpretable metric for team quality.