Phil Steele Blog • July 17, 2015

 

 

How Often do Underdogs Win Outright?

 


One aspect of college football that I like to monitor is a team’s performance relative to expectations.  As most of you know, I pride myself on producing the most accurate college football magazine and I rely heavily on data analysis using computers to arrive at accurate projections.
Upsets are one of the most exciting aspects of college football and they occur on a weekly basis.  Of course, upsets come in different shapes and sizes, which makes it useful to separate this information into different categories.  By looking back to 1997, we can benefit from having a significant number of data points (i.e., almost 26,000 games) from which to draw conclusions.

Beginning with the large point spreads, you will see that Las Vegas does its job well picking out big favorites. Since 1997 there have been 783 teams that have been favored by over 31.5 points in a game.  In those games, only SEVEN have lost the game straight up, including zero last year. The largest all time upset took place in 2007 when Jim Harbaugh led Stanford (+41) to an upset victory at USC.  In addition, the second biggest upset took place in 1997 when Syracuse (+37) stunned Louisville. The other 5 huge upsets since 1997 were: Texas State (+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. In sum, if your team is installed as an underdog of 31.5 points or more in a game, they basically have a 0.89% chance of winning. 

Now let’s look at the chances of your favorite team winning if they are installed as an underdog of 24.5 to 31 points. Since 1997, there have been 942 teams that have been favored by 24.5 to 31 points in a game. Naturally, the underdogs are more likely to spring an upset within this data set.  During our relevant time span, there have been 36 upsets in this range of point spread, or roughly 2 per year, for a 3.8% upset percentage.  While 2013 was a bit of an outlier with 4 such upsets, there were zero upsets by a 24.5+ underdog in 2014. 

Again dating back to 1997, there have been 1,620 games that fall into the 17 to 24.5 point category. During this time period, only 105 underdogs have won outright, which works out to a modest 6.48% winning percentage. Last year, there were eight such upsets in this category, the largest of which was Arizona’s stunning 31-24 victory over Oregon at Autzen Stadium. Oregon, of course, exacted a measure of revenge in the Pac 12 Championship with a dominant win over the Wildcats.

Since the majority of games are somewhat evenly matched, I narrow the point differential for each grouping as the point spreads narrow.  Thus, my next category includes teams that are favored by 14.5 (i.e., more than 2 TDs) to 17 points, of which there have been 1,050 during our selected time span. Within this category, 134 of the underdogs have won straight up, so the upset rate roughly doubles to 12.8%. Seven teams pulled an upset from this category in 2014, led by Boston College running wild over USC as a 17 point home underdog.

The next category ranges from 10.5 to 14 points. From 1997 – 2014, there were 1,745 teams that were favored by this amount and 364 upsets, which was 20.9% of the time. That still means that roughly 4 out of every 5 teams favored by 10.5-14 points win the game, with an upset occurring once every 4.8 times. 
My next tier looks at point spreads that were greater than a touchdown, so it includes lines between 7.5 and 10.  In our selected time period, there were 1,657 teams that were favored by 7.5 to 10 points and 434 upsets, which translates to a 26.2% upset rate (i.e., greater than a 1 in 4 chance).

As we move closer to toss up territory, the volume of games and the percentage of upsets both increase.  From 1997-2014, there were 3,044 teams that were favored by more than a field goal but not more than a touchdown (i.e., 3.5 to 7). In those 3,044 teams, underdogs have claimed victory in 1,069 games for a percentage of 35.1%.  Last year, the upset rate was a bit higher with 38% in this category.

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.  Since 1997, there have been 2,020 games in this category and 947 upsets, so the underdogs are not quite a 50-50 proposition at a 46.9% winning percentage.

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


1997-2014

Favorite of

# of Games

Lost Outright

%

31.5 or more pts

783

7

0.89%

24.5-31

942

36

3.82%

17.5-24

1620

105

6.48%

14.5-17

1049

134

12.77%

10.5-14

1745

364

20.86%

7.5-10

1657

434

26.19%

3.5-7

3044

1069

35.11%

3 or less

2020

947

46.88%

Finally, among the media and fans, there is a popular perception that parity is rampant in college football compared to years past, which has led to a rash of wild upsets.  To test this narrative, I have compiled a chart of data from the last 6 years as opposed to the last 18 years in our main chart.  As indicated below, there have actually been fewer upsets in the last 6 years than in the last 18 years, which means that upsets have been declining slightly.  You can see that the discrepancies are not major and I am not calling this a trend, but merely pointing out that there is no evidence of rising upsets in college football in the past few years.


2009-2014

Favorite of

# of Games

Lost Outright

%

% of Change

31.5 or more pts

411

2

0.49%

-0.40%

24.5-31

355

13

3.66%

-0.15%

17.5-24

607

35

5.76%

-0.72%

14.5-17

399

46

11.52%

-1.25%

10.5-14

600

123

20.50%

-0.36%

7.5-10

601

158

26.28%

0.09%

3.5-7

1113

410

36.83%

1.72%

3 or less

745

322

43.22%

-3.66%

 

 

 

2015 College Football Preview
available right now on iTunes & Google Play!



1). It is available WEEKS before the magazine hits the newsstands on July 1st! It not only allows you to get the magazine weeks before it hits the book stores but it will save you wear and tear on your printed magazine. If you get both you can use it as your mobile version of the magazine and that will keep your actual magazine from getting dog eared early in the year.

2). It’s easy to use and works perfectly for both the iPhone, iPad and Android Devices. The download process may take up to 10-15 minutes but that’s because we have 352 pages with jampacked information!

3). No bugs, no issues, take it with you where ever you go!

4). Non-U.S. Customers: The magazine will not be on the newsstands internationally, the only way to get the 2015 College Football Preview is to order through our offices and pay at least $35 (includes S&H) or get it on iTunes or Google Play  for just $9.99.

5). It’s only $9.99! On the newsstand, the magazine will cost you $12.99.

Please note: after downloading and purchasing the App, please make sure you search for 2015 if the old 2014 cover/edition comes up on your screen.

Don’t wait three weeks until the magazine hits the newsstands. 

WE HAVE ALSO RECEIVED OUR SHIPMENT IN THE OFFICE! Don’t wait until July 1st for the 2015 Phil Steele College Football Preview to hit the newsstands order it now and we’ll ship it out within one business day. Call 1-866-918-7711. Or order online at the Phil Steele Store.

The first magazine is $12.99 plus $6 for Priority Mail service. Add a second for $10 if delivered to the same address at the same time. OR, take advantage of our 7-6-5 special.

Seven 2015 College FB Previews delivered to one address just $65 (and that includes postage!)