The Bigger the Favorite, the Worse the Cover

Methodology

We bucketed every game by the absolute closing spread (spread_favorite) and computed the favorite's cover rate (favorite_spread_outcome) in each band, for the full history and again for 2001–2025.

import pandas as pd
df = pd.read_csv("spreadspoke_enhanced.csv", low_memory=False)
df['absp'] = df.spread_favorite.abs()
for lo, hi in [(0,2.5),(3,3),(3.5,6.5),(7,9.5),(10,13.5),(14,40)]:
    s = df[(df.absp>=lo)&(df.absp<=hi)]
    fc, fn = (s.favorite_spread_outcome=='Cover').sum(), (s.favorite_spread_outcome=='No Cover').sum()
    print(f"spread {lo}-{hi}: fav cover% {100*fc/(fc+fn):.1f} (n={fc+fn})")

The Analysis

Intuition says the best teams should bury overmatched opponents and waltz past the number. The data says the opposite: the heavier the favorite, the worse the cover rate. Teams laying two touchdowns or more have covered roughly 45% of the time — the weakest band on the board — while pick'em-range favorites hover right around 50%. The cover rate slides downhill almost monotonically as the spread grows.

The mechanics are familiar to anyone who has watched a 31-3 game crawl to the finish. Big favorites pull their starters, stop calling aggressive plays, and let the backups bleed clock, while the trailing team racks up cosmetic points in garbage time — a classic backdoor cover.

This isn't a quirk of the 1970s, either: the modern-era numbers tell the same story. The reader's rule of thumb is to be skeptical of the “obvious” blowout — the market already expects it, and history says the big favorite cashes a ticket less than half the time.