NFL Favourites by Week: Do They Win and Cover More Late in the Season?

Every September I watch the market price favourites as if last year's rosters had simply carried over. The question here is whether the week of the season changes how favourites perform against their moneyline price and against the closing spread. The study covers 5,199 regular-season games with closing lines, split into five week bands.

The headline answer is yes, and the shape is clear. In Weeks 1-4 favourites won 62.6% of games while priced at 65.2%, and covered only 46.5%. By Weeks 17-18 they won 73.7% and returned +2.6% on the moneyline. I use this calendar as a check on everything that goes onto my free NFL predictions page.

Window 1Early favourites win less often than their price

Each row sets the margin-free moneyline chance against how often favourites actually won. Weeks 1-4 is the only band where the gap clearly favours the underdog: priced at 65.2%, won 62.6%. Through the middle of the season the two dots sit close together, at 66.5% priced and 66.3% won in Weeks 5-8, then 66.7% and 67.9% in Weeks 9-12.

Then Weeks 17-18 flips the picture. Favourites were priced at 69.7% and won 73.7%, the widest gap in the chart, and it runs in the favourite's direction.

The September miss fits what I see on film. Offensive lines haven't gelled, new coordinators are still showing their hands, and the market leans on reputation. That uncertainty costs the side priced to win more than the side priced to lose.

Window 2Moneyline returns swing from underdogs to favourites

Flat-staking every side tells the same story in money. In Weeks 1-4 backing every favourite returned −7.3% while every underdog returned +1.5%, the only positive underdog figure on the chart. Across all weeks both blanket approaches lose, with favourites at −3.2% and underdogs at −3.4%, which is roughly what the margin alone would cost you.

Weeks 17-18 is where it gets dramatic. Favourites returned +2.6% and underdogs collapsed to −18.8%.

Here is my main caveat. The final fortnight holds only 395 games, and those results are shaped by teams resting starters or playing out the string once they're eliminated. That underdog figure reflects rosters as much as prices, so I read the late-season favourite return as a reason to check team news hard, not as a standing edge.

Window 3Favourite cover rate climbs steadily through the season

Against the closing spread, the early weeks stand out again. Favourites covered just 46.5% in Weeks 1-4, then 49.2%, 49.1% and 49.6% through the middle bands, reaching 50.8% in Weeks 17-18. Across the whole sample they covered 48.8%.

Keep the break-even mark in mind. At standard −110 juice you need 52.4% to profit on a spread bet, and no band of favourites gets there.

So the late climb is not a reason to lay points blindly. The useful signal sits at the other end: September is the stretch where taking the points has historically held up best, and favourites show up weakest against the number.

Window 4Short September favourites cover least of all

Splitting by spread size shows where the early-season weakness is concentrated. Favourites of up to 3 points covered only 43.7% in Weeks 1-4, against 49.1% in Weeks 17-18. Favourites of 3.5 to 7 points covered 48.1% early and 50.4% in Weeks 13-16, while the 7.5-plus group went from 48.6% early to 51.2% at the end.

These are matchup games where a field goal decides things, and early in the year the market seems to give too much credit to the stronger name. A short home favourite in Week 2 is often a team the public remembers, not the one on the field now.

Each cell holds only a few hundred games at most, so I don't chase individual squares. A difference of a few points between neighbours is noise. The short-favourite September figure is the one I treat as a pattern, because it lines up with the moneyline and cover-rate charts.

Written byVanessa Brooks

I'm Vanessa Brooks, and I write the American football predictions at fixedmatches.vip — breaking down spreads, totals and the matchups that settle them.

View profile →