🏈 College Football Totals: Pick the Side, Not the Number

Since 2001 I have bet fewer than 200 college football OVERs. In that same stretch I have bet more than 1,700 UNDERs.

That is not an accident. I have known why the UNDER is the better side for most of my career, and I have been telling people so for years. What I had never done was measure how much that instinct was worth. So I went and measured it.

The first half of what came back is the answer I expected, and it is bigger than I would have guessed. The second half took a theory I had carried for a long time and killed it.

🏈 Every college football pick I've made is posted with its result attached. See the full archive.

📉 The OVER and the UNDER are not the same bet

Every market figure in this article is what a blind bet returned. Blind means backing that side every single time it came up across twenty seasons, with no handicapping applied and no opinion about the two teams.

I pulled every college football game with a posted total going back to 2006, 17,511 of them. Here is what the two sides paid:

The two tickets cost the same and pay the same, and the only thing separating them is which word you circled. That word is worth a little more than four points of return.

That gap is a large part of why I have a career in this sport.

🎉 Why the UNDER is the cheaper side

Nobody at a sportsbook decided to make UNDERs cheap. The price is a record of where the money already wants to go, and in a totals market the money wants to go one way.

Bettors are fans. A fan sits down on a Saturday to watch football, and what he wants to watch is scoring. Nobody puts a ticket in his pocket hoping to spend three hours rooting for punts and missed field goals.

The OVER is the fun bet. It gives you something to cheer on every drive and it stays alive into the fourth quarter. The UNDER asks you to want less football, and almost nobody wants less football.

That preference has to go somewhere, and where it goes is the price. It has been going there every Saturday for twenty seasons.

I have been saying this for years, long before I could put a figure on it. UNDERs are not fun and they are not sexy. They win more often. I would rather win.

“Nobody puts a ticket in his pocket hoping to spend three hours rooting for punts.”

🧮 What "four points" actually costs you

At the standard price of −110 you have to win 52.38% of your bets to break even. That figure is arithmetic. It does not care which sport you pick or how good you are.

What changes is where you start from before you have done any work.

Back every college football UNDER blind and you win 51.07%. To get to break-even from there you need to be right about 1.3 more games in every hundred than that blind bettor.

Back every OVER blind and you win 48.93%. Now you need 3.5 more games in every hundred.

I have written elsewhere about what a strong professional handicapper is actually worth, and my answer is two to five extra wins per hundred more than a blind bettor in the same spot. Hold that against the two rows above. The UNDER asks for a fraction of what a good handicapper has. The OVER asks for most of it, before he has beaten anybody.

You can spend a decade getting better at reading college football offenses, or you can stop betting OVERs. One of those is free.

🤔 Then which totals should you be attacking?

Everybody who bets totals has a theory about the size of the number, and for most of my career I had one too.

High-scoring games run away and land OVER. Big totals are inflated by public money on shootouts and should be faded. Low totals are traps because one broken tackle ruins them. I have heard all of these and tested versions of them, and the assumption sitting underneath every one is that the number on the board is telling you something.

So I split all 17,511 games by the size of the total. Six bands, from games totaled in the low forties to games totaled near seventy.

📊 Six bands, and your eye will lie to you

Look at that UNDER column and you will find a pattern in it. The two smallest totals look like the place to be, and the middle band looks like the one to avoid. That is a system, sitting right there, and it is worth roughly two points.

It is not there, and this is the part worth understanding, because it is the mistake I have watched bettors make with every stat sheet ever printed.

Each of those band figures carries a margin of error of about three to four and a half points either side. The distance from the best band to the worst one is 2.2 points. The uncertainty on any single row is wider than the entire spread between all six of them. Run the proper test across the six and the result comes back as close to "no pattern whatsoever" as a statistical test can land.

Six numbers that are genuinely the same will still print as six different numbers. One of them has to be the highest. That does not make it the best band, it makes it the top of a flat line.

🕐 The season the rules changed

In 2023 college football changed how the game clock runs. Fewer stopped clocks meant fewer plays, and the scoring came down with them. The average total in my own database fell from 54.6 in 2022 to 52.0 in 2023, and it has stayed down.

That raises a fair question about everything above. A twenty-season average is not worth much if the seasons since the rule change broke it.

I can answer it, because my database keeps two numbers on every game. There is the total the book first hangs, days before kickoff, and there is the total it closes at once the money has come in. Those two are usually not the same, and the difference between them is bettors moving the number.

Measured against the opening total, across 13,742 games going back to 2007, the OVER came in 48.23% of the time before 2023 and 48.22% of the time after.

That is the same figure twice. The rule change did not make college football teams beat the posted number any more often, because the oddsmakers hanging those numbers had read about the rule change too.

What moved was everything that happened after the total went up on the board. Before 2023 it drifted about a tenth of a point between the opener and kickoff. In 2023 it fell almost a full point on average, and on more than half of all games played the total came down by a full point or more.

Everybody betting college football knew scoring was down, and the money went to the UNDER. How much of it came from recreational bettors and how much from professionals, I cannot tell you. Professionals move a total further and faster than the public does, which makes them the likelier driver, but I have no record of who was betting which side on those games and I will not pretend otherwise.

Either way it is not the preference I described earlier. Fans wanting to watch points is the standing condition that pays for the four points, and that has not gone anywhere. This was a rule change being priced in, and for two seasons the number went past where it needed to go. In 2025 the blind UNDER graded off the opening total turned a profit again.

One thing has not gone back. A total of 55 used to sit right around the middle of a Saturday board. Since 2023 it sits above two thirds of it. The same number now describes a different kind of game than it did before the rule change, which is one more reason not to read anything into the number by itself.

🎯 What that means for the number on your screen this Saturday

Two findings, and they point in different directions, so keep them apart.

The side matters. UNDERs beat OVERs by four points of return across 17,511 games. The reason has not changed in twenty seasons, because what bettors want to watch on a Saturday hasn't either. The two seasons after the 2023 rule change are the one stretch where that gap closed, and by 2025 it was back.

The size does not. A 68 and a 43 are the same bet with different digits on them. If your process starts by scanning the board for big totals or small ones, that step is doing nothing for you.

None of that says totals are unbeatable, and none of it says handicapping a total is pointless. It says the shortcut everybody reaches for first has no measurable value in this sport. Whatever edge exists in a college football total comes from the teams on the field and the conditions they are playing in. It does not come from the size of the line.

That is a harder job than the shortcut, and it is the only one of the two that pays.

💵 My own record, posted and scored

Since the 2013 season I have gone 891-781-34 on college football totals. That is 53.3% of the ones that decided, and every one of them is posted with its result attached, the same as everything else I have sold since 2001.

Set that against the blind UNDER bettor's 51.07% and it is 2.2 more wins per hundred. It sits inside the two-to-five range I said a strong handicapper is worth, which is where I would expect an honest number to land and nowhere near where a tout would put one.

In money it came to 49.57 units across thirteen seasons, betting between one and two units a game. Some of those seasons lost, which is what a real edge looks like from the inside. Anybody promising you it looks smoother than that is selling you something.

🎯 Stop reading the number, start reading the game

Take the UNDER seriously and take the OVER rarely. That is not a system and I am not handing you one, but it is where the price sits before either of us has an opinion about a football game.

Stop sorting by the size of the total.

And do the actual work on the games, because that is the only part of this still worth anything. The four points on the UNDER side are available to everybody reading this, which means they are not an edge, they are a starting position. What you build on top of them is the edge.

There is a question sitting underneath all of it, and the betting public answers it every Saturday afternoon. Are you in this for fun, or are you in it to make money?

Twenty seasons of prices say the public is mostly in it for the fun, and the fun is on the OVER. That preference is where the four points come from, and they are there for anyone willing to sit on the other side of it and root for punts.

I put out a complimentary pick every day, and if you want my college football card for the season, the packages are here. Either way, the numbers above do not change.

The blind-bet figures here come from 17,511 college football games played since 2006, graded at the standard −110 price. The opening-total figures come from my own database, which carries both the opening and the closing number on 13,742 games since 2007. Lines move; confirm the current number before you bet it.