Home Field Is Dying, and Wind Kills Overs
I graded 8,000+ NFL games against the closing spread, going back to 1995. Three things jumped out, and one of them quietly rewrites a betting rule everyone still repeats.
1. Home field isn't what it used to be
Every old betting guide says "give the home team 3 points." That used to be right. It isn't anymore. Here's the average home margin by decade:
Non-neutral-site games only. Live from our database.
The drop is steady and it's big. Home teams used to win by 3.3 in the 90s; now it's about 1.9. Nobody's totally sure why — better travel, less crowd intimidation on replay-reviewed calls, quieter pandemic seasons that never fully bounced back — but the number is what it is. If your mental model still spots the home team a field goal, you're giving away about a point and a half on every game.
2. Does the spread actually predict who wins?
Yes, almost exactly. If you line up every game by how big the favorite was and check how often they won outright, the real win rate tracks a clean bell-curve prediction the whole way up:
Spreads with 40+ games. Hover a dot for the exact rate and sample size.
The spread is a genuinely good estimate of the real margin. The spread of the final scores around that line — how far blowouts swing from nail-biters — works out to about 13 points. That number is exactly what a fair model needs to price the line, and it's why our own model uses it.
3. There's no free money lying around
People love "just bet home dogs" or "fade the public favorite." The data says none of the easy rules work. Against the spread, favorites covered 47% of the time and home teams 48% — both just under breakeven once you pay the vig. That's the signature of a sharp market: the easy edges are already priced out.
What IS true is the key numbers. Final margins cluster hard on 3 and 7, because games are decided by field goals and touchdowns:
This is why −2.5 and −3.5 are very different bets: the 3 sits between them.
4. The only weather that matters: wind
Everyone blames the cold for low-scoring games. The data says cold barely matters on its own — wind is the real killer. Look at the over rate and the average total in calm games vs windy ones:
Games with a posted total and a wind reading.
Calm games go over about half the time, like you'd expect. But once the wind hits 15 mph or more, the over rate falls to about 41% and the average game scores around 39 points instead of 45. At 20+ mph it drops to roughly 37. Kicking, deep passing, and timing all fall apart in the wind. Cold by itself? Mostly a red herring — cold games are usually calm cold games, and those score fine.
The honest caveats
- It's historical. 30 years of closing lines. The market adapts; the home-field number is literally proof that things drift.
- Single closing line, no shopping. The market-efficiency numbers use one closing spread. Shopping the best number across books is its own (real) edge on top.
- This is descriptive, not a system. "Home field shrank" and "wind kills overs" are facts about the past, not guaranteed money. The books know all of this too.