NFL Totals by Month: Do December and January Games Go Under?

Every winter I hear the same line: it's December, it's cold, take the under. I wanted to test it properly, so this study covers 5,199 NFL regular-season games from 2006 to 2025 with closing totals. It asks whether the month changes how often the over lands, and whether open-air stadiums behave differently from domes.

The short answer is that the calendar barely moves the needle. Across the whole sample the over hit 49.8%, and no month strays far from that. The lowest month is November, not December, and the reason the late-season under fails is that the market has already shaded the total down. Treat this as background for the matchup work on my American football predictions for today page, not as a replacement for it.

Window 1No month breaks far from a coin flip

This table carries the headline. September sits at 50.3%, October at 50.7%, November at 48.6%, December at 49.6% and January at 50.5%. The 95% ranges are the part that matters most. They run ±2.8 or ±3.0 for the big months and ±6.9 for January, which means most of these months overlap each other comfortably.

November is the softest month for the over. Its fair under price of 1.95 still sits well short of the −110 break-even of 52.4%. Backing the under every November week would not get you past the vig.

I read this table as a warning against blanket monthly angles. Gaps of a point or two across samples of this size are mostly noise. The honest caveat for the whole study belongs here: weather is never measured directly, so the stadium split later on only stands in for how exposed a game is to the elements.

Window 2Bookmakers trim the late-season total before kick-off

This is why the December under fails. The average closing total drops from 44.7 in November to 44.0 in December and 43.2 in January. Scoring falls far less: 44.9, 44.8 and 44.2 points across the same months. The market expects the cold to bite, and it shades the number down further than the scoring actually falls.

In every month, average points scored sit slightly above the average line. October shows the widest gap, with a 44.6 total against 45.4 points scored.

For me the lesson is simple. The narrative about winter football is already in the price. If you want an edge late in the year, it has to come from a specific matchup or a specific weather report, not from the date on the calendar.

Window 3Domes run warmer for the over outside December

Splitting by roof is my best proxy for weather exposure. Open-air games stay close to the coin flip all season, bottoming out at 47.5% in November. Open-air December, the one game type everyone wants to back under, went over 49.8% of the time.

Domes and closed roofs lean over in most months: 51.0% in September, 54.2% in October and 54.5% in January. December is the exception at 49.0%. The January dome figure rests on a small sample, so I would not build anything on it alone.

What I take from this is that the indoor over is the more interesting thread, not the outdoor under. Even so, it is a lean worth checking against the matchup, not a rule.

Window 4The monthly pattern flipped between the two decades

This chart is the reason I distrust any seasonal rule. From 2006 to 2015 the over rate faded through the year, from 51.9% in September to 48.8% in December and lower still in January. That era built the December-under reputation.

From 2016 to 2025 the shape reversed. The over started at 48.7% in September and climbed to 50.4% in December, with January higher again. The old pattern did not just weaken; it turned upside down, and the market adjusted to it.

When a calendar edge flips sign from one decade to the next, I treat it as something the market adapts to, not a law of football. Patience beats an angle that the bookmakers have already learnt.

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 →