The Key-Number Tax: Why 3 and 7 Rule NFL Spreads

Methodology

We counted the frequency of every winning margin (margin_of_victory) across all decided games, then measured how often games landed exactly on the spread (a push) when the closing number (spread_favorite) was a key value like 3 or 7, using favorite_spread_outcome.

import pandas as pd
df = pd.read_csv("spreadspoke_enhanced.csv", low_memory=False)
mov = df.margin_of_victory.dropna(); mov = mov[mov>0]
print({k: round(100*(mov==k).mean(),1) for k in [3,7,6,10]})
for k in [3,7,10]:
    s = df[df.spread_favorite.abs()==k]
    print(f"spread {k}: push {100*(s.favorite_spread_outcome=='Push').mean():.1f}% (n={len(s)})")

The Analysis

NFL scoring is built out of threes and sevens — field goals and touchdowns — so the final margins inherit that rhythm. Three points is the single most common margin in league history, deciding more than one in seven games. Seven is next. Everything else trails well behind. That clustering is the entire reason “key numbers” exist and why so much line movement happens around them.

For anyone reading a number, the practical consequence is push risk. A spread sitting exactly on 3 lands on the number — a push, your stake returned, no win — about 9% of the time. That's why books charge a premium to move a line from -3 to -2.5, and why that particular half-point is worth more than any other on the board. Getting off 3 (or onto it, if you're laying) materially changes how often you win.

You don't need a model to use this. Just respect the key numbers: half-points around 3 and 7 are expensive for a reason, and the chart below shows exactly why.