How often to underdogs win outright?

Naturally there are different levels of underdogs and if you know me by now, you know that I will head to my computer and get all the details. There has been a lot of college football played since 1997, 26-years’ worth of games.

Let’s look at the biggest favorites and you will see that Las Vegas does its job well picking out big favorites. Since 1997 there have been 892 teams that have been favored by more than 35.5 points in a game. Of those 892 only SEVEN have lost the game straight up. The biggest upset ever was in 2017 when Cam Newton’s little brother guided Howard +45 to and outright upset of UNLV. The biggest upset involving Power 5 teams was when Stanford (+41) upset USC in 2007. In 2015 The Citadel were +38 and upset South Carolina for the 3rd biggest upset ever. 2007 featured the 2nd biggest Power 5 upset when Syracuse (+37) beat Louisville. The other 4 huge upsets since 1997 were: Texas St (+36.5) over Houston in 2012, Temple (+35.5) over Virginia Tech in 1998, Central Michigan (+35.5) over Western Michigan in 2000, James Madison (+35) over Virginia Tech in 2010 and North Texas (+32) over Texas Tech in 1997. Basically, if your team is installed as an underdog of 35.5 points or more in a game, they have a 0.00785% chance of winning.

Now let’s look at the chances of a team winning if they are installed as an underdog of 24.5 to 35 points. Since 1997 there have been 2016 teams that have been favored by 24.5 to 35 points in a game. You would expect the underdogs would have a better chance of winning outright than the above teams and naturally they do. In that span 63 teams that have been favored by that margin have lost straight up. In 2020 the biggest upset was Kansas St +27.5 upsetting Oklahoma 38-35. Maryland was +27 but pulled an upset of Penn State. In 2021 ULM was +33 and upset Liberty. Bowling Green was +31 and upset Minnesota. Kansas was +31 and upset Texas. Those were the 3 biggest upsets. Last year there were two upsets by teams +24.5 or more. Middle Tennessee was +24.5 and upset Miami Florida and New Mexico St was +24.5 and upset Liberty. The number of upsets did not go up as much as I thought they would as upsets occurred only 3.125% of the time.

You would figure that the percentage of upsets in the 17.5-to-24-point favorites level would go up drastically from that 3.1% in the previous category. Since 1997 there have been 2,464 teams that have been favored by 17.5 to 24 points. I am still a little surprised that there were only 182 upsets in this level of favorites over the 26-year span. That is just 7.39% upsets, which comes out to 1 upset loss for every 13.5 teams that are favored by 17.5 to 24 points. In 2020 there were six such upsets. LSU was +23.5 and thanks in part to a shoe toss upset Florida. Michigan St was +22 but took down rival Michigan. Rice was +22 and stunned a ranked undefeated Marshall on the road. Tulsa was +20.5 but upset UCF for a 2nd straight year. Maryland was the first team in FBS history (established 1979) to win consecutive games as a dog of 18+ when they upset Minnesota as a 20.5 dog then Penn St listed above the next week. The other was when New Mexico +18 took down Wyoming.      In 2021 there were nine upsets by teams that were in the +17.5 to +24 range. Illinois was +24 at Penn St, ULM was +23.5 at home vs Troy, Rice was +23.5 at UAB, South Carolina was +20.5 at home vs Florida, New Mexico was +20 at Wyoming, Northern Illinois was +19 at Georgia Tech, Texas A&M was +18 at home vs Bama, Utah St was +17.5 at Wash St and Florida St was +17.5 at North Carolina (can you believe that was the spread just two years ago).    Last year the number rose to 13 upsets in this category with the biggest upsets being Georgia Southern +23.5 over Nebraska, Connecticut +23 over Fresno, Georgia Tech +22 over Pitt, South Carolina +22 over Tennessee, Georgia Tech +21 over North Carolina, Eastern Michigan +20.5 over Arizona St and Marshall +20.5 over Notre Dame.

I will start breaking it down into smaller categories as we continue to go down in level of favoritism. The next category to look at are teams that are favored by 14.5 (more than 2 TD’s) to 17 points. Since 1997 there have been 1586 teams that have been favored by 14.5 to 17 points in a game. Of the 1586 teams favored by this amount 213 have lost the game outright. That is almost double the odds as the category above as 13.4% of the teams favored by 14.5 to 17 points lose outright and an upset occurs 1 time for every 7.44 games played in that level. There were 19 such upsets last year with the biggest being Bowling Green +17 upsetting Marshall the week after Marshall upset Notre Dame.

The next level down that I will look at is the 10.5-to-14-point category. From 1997–2022 there were 2,657 teams that were favored by this level. Of those 2,657 games there were 538 upsets, which was 20.2% of the time. That still means that roughly 4 out of every 5 teams favored by 10.5-14 points wins the game with an upset occurring once every 4.93 times.

From 1997-2022 there were 2,412 teams that were favored by 7.5 to 10 points and of those 2,412 teams there were 627 upsets. That means that 26.0% of the time a team was favored by 7.5 to 10 points they lost outright and that translates into 1 upset every 3.84 games.

Since 1997 there have been 4,623 teams that have been favored from 3.5 to 7 points. These are teams that are expected to win the game, but the game should be close. The last time we tightened up the category the upsets went from every 4.93 games to every 3.84 games. Should we expect 1 in every 3 games here? Since 1997, 1632 of the 4,623 teams that were underdogs of 3.5 to 7 points have pulled outright upsets and that translates into 1 every 2.83 games or 35.3% of the time.

Now we get to the category that will tell us if Las Vegas knows what they are doing. Games where a team is favored by 3 points or less. These games are basically toss-ups, but Vegas favors one team and makes them the favorite. What percentage of teams pull minor upsets as underdogs of 3 points or less? As I type this, I do not know the answer, but I will guess 45% of them. Now let’s go to the computer. Since 1997 there have been 2,958 favorites of 3 points or less. My 45% guess would have 1331 minor “upsets”. The actual answer is 1410 “upsets” by teams that are underdogs of 3 points or less which comes out to be 47.7% and that shows these games are definitely toss-ups!

How about when games are pick ‘em or even. This means that the visitor is regarded as the stronger team but not by much. I guess this will determine who much home field means. I will ignore the neutral site E games. There have been 127 such E games which have been on one teams home field since 1997. The visitor (stronger team) is 69-58 54.3% in these games.

Here is a quick chart showing the chances of a team winning outright depending on where Las Vegas sets the line.

Now lets take a look at the last 12 years and see if the numbers remain similar. Here are the numbers from the last 12 years.

As you can see the numbers continue to remain consistent and not number was off than more than 1% so we are not trending in any direction just staying steady so this year’s upset numbers should remain consistent.

Remember the Digital Magazine is updated thru the month of September with circles around players who are out for the year and changes like Northwestern head coach is now David Braun being updated. To get the digital version right now CLICK HERE.


Get Phil’s Extended data and Plus your odds. Click on Learn more to see exactly how Phil uses Phil Steele Plus. FCS only access also available.