clock menu more-arrow no yes mobile

Filed under:

Reviewing Pac-12 Game Predictions

Several sites had game-by-game predictions before the season. How well did they do?

Notre Dame v Stanford Photo by David Madison/Getty Images

Back in early August, I published an article which reviewed several sites which had game-by-game predictions for all of the Pac-12 games. This will review those to see both which sites did the best and which teams were the easiest and hardest to predict.

Here are the sites that I tracked with links to their predictions.

I’m also including ESPN’s FPI which had preseason predictions (win percentages) for each game.

Game-By-Game Predictions

Here is a look at how these sources did with picking each of the games.

Preseason Accuracy Review

Source Correct Wrong %
Source Correct Wrong %
CBS 68 22 76%
College Football News 64 26 71%
College Football with Sam 71 19 79%
Gridiron Expert 69 21 77%
SG1 Sports 65 21 76%
Lucas Ross Sports 67 23 74%
All Sports Central 62 22 74%
ESPN FPI 64 26 71%
UWdadVanc 71 19 79%

I’m sure that it won’t shock many fans, but ESPN’s FPI had the worst predictions of all of the ones that I tracked. They got just 71% correct. They were also tied with College Football News.

The best predictions were from College Football with Sam at 79%. While that seems good, I was also at 79% with my preseason predictions, so I’m not sure that I would trust any of these other sources over my own instincts and analysis.

I will note that ESPN’s FPI updated their predictions each week. If you look at the accuracy of the predictions for each week’s games, they were at 79%. That is an improvement-but only to the best predictions before the season.

Total Wins

Picking individual games before the season starts is understandably difficult. But how did these sources do in predicting the win totals for each team.

Win Total Accuracy

Team Actual Wins CBS CFN College Football with Sam Gridiron Expert SG1 LucasRoss All Sports Central ESPN FPI UWdadVanc
Team Actual Wins CBS CFN College Football with Sam Gridiron Expert SG1 LucasRoss All Sports Central ESPN FPI UWdadVanc
Arizona 9 6 6 6 5 6 5 * 5.6 5
Arizona State 3 3 5 1 4 4 3 6 5 4
California 6 4 5 4 5 * 4 * 5.7 5
Colorado 4 3 3 3 3 4 3 * 2.8 3
Oregon 11 9 9 10 10 10 9 10 8.9 9
Oregon State 8 10 8 10 9 8 10 8 8.6 10
Stanford 3 4 2 2 1 * 2 * 4.3 2
UCLA 7 9 8 8 8 8 8 8 7.5 9
USC 7 11 11 10 11 10 10 11 10 9
Utah 8 10 10 10 10 9 8 10 8.5 9
Washington 12 11 9 11 10 10 9 11 7.7 10
WSU 5 4 6 6 6 6 9 * 5.6 6

*-No win total projected (based on the game-by-game predictions), or game predictions were inconsistent.

Arizona exceeded the win total expectation of every one of the sources (including me). Similarly, USC failed to achieve the win total expectation of every source (also including me-although I was closer than any of the others).

There were 4 other cases where the win total expectation was off by more than 2 wins (up or down). All Sports Central had Arizona State getting bowl-eligible (6 wins), and they clearly did not do that. ESPN’s FPI had Oregon winning 8.9 games and they won 11. Lucas Ross Sports had WSU winning 9 games when they only won 5. And finally, several places had UW winning less than 10 games. One was College Football News (CFN). The next was Lucas Ross Sports. And the single worst win total projection of all was ESPN’s FPI which had UW only winning 7.7. (I’m sure that there are many Husky fans who are not surprised the ESPN’s FPI was the worst.)

Here’s a look at how often each source was off by more than 2 wins:

  • CBS Sports: 2 of the 12 teams.
  • College Football News: 3 of the 12 teams.
  • College Football with Sam: 2 of the 12 teams.
  • Gridiron Expert: 2 of the 12 teams.
  • SG1 Sports: 2 wins for 2 of the 10 teams that they had game-by-game win predictions.
  • Lucas Ross Sports: 4 of the 12 teams.
  • All Sports Central: 3 of the 9 teams that they had game-by-game win predictions.
  • ESPN’s FPI: 4 of the 12 teams.

I was only off by more than 2 wins for just 1 of the Pac-12 teams (Arizona). So again, I’m not sure that I would trust any of these sources over my own instincts and analysis.

Team Accuracy

To see which teams were easiest and hardest, I combined the win percentages from each of the sources. Here are the results.

  • Oregon: 84%
  • California: 82%
  • Washington: 81%
  • Stanford: 76%
  • Oregon State: 74%
  • Arizona: 71%
  • UCLA: 71%
  • Arizona State: 70%
  • USC: 70%
  • Utah: 68%
  • Colorado: 67%
  • WSU: 67%

Oregon and Washington were considered two of the top teams in the conference before the season started, and the fact that they have lived up to that explains why they are in the top 3. California was supposed to be a fringe bowl team; and that’s what happened-making a bowl game with their last game. Stanford upset a couple of teams, lost a game they were supposed to win, but otherwise did what was expected-lose most of their games.

Despite Arizona exceeding expectations, they are in the top half. That’s because the sources were (consistent and) perfect with their predictions for the first half of the season.

It’s probably not surprising that Colorado is at the bottom. But that has to do more with their wins at the beginning of the season than their losses after that. WSU being at the bottom had a lot to do with their 6-straight losses, but also their wins over Wisconsin and Oregon State.

Individually, each of the sources had a less than 67% accuracy for at least 1 team-except for the Gridiron Expert. They had two at 92% and 4 at 67%. ESPN’s FPI was less than 67% for 3 teams, and they were above 75% for only 2 teams (Cal and Stanford). College Football with Sam was above 75% for 10 of the 12 teams, but only 58% for the other two teams (UCLA and WSU). College Football News wasn’t above 75% for any team and at 58% for 2 teams (Colorado and WSU).

With my personal predictions, I was almost perfect for Cal. The only game that I missed was their win over UCLA. And I was above 80% for 5 On the other end, I did not expect Arizona to even make a bowl game. And I also didn’t expect UCLA to lose 3 of their last 4-and win the one that they did! I was at 67% for both teams.

Poll Results

With that article back in August, I included a poll which asked which team was most likely to exceed its maximum win total expectation. Four of the teams did exceed their maximum win total expectation.

  • Arizona’s maximum was 6. They exceeded it by the most with 9 wins. 17% of the respondents picked Arizona.
  • California’s maximum was 5 and they won 6. 5% of the respondents picked Cal.
  • Oregon’s maximum was 10 and they won 11. Only 4% of the respondents picked Oregon.
  • UW’s maximum was 11 and they, of course, won 12. Almost a quarter (24%) of the respondents picked UW.

Kudos to those who picked any of those 4 teams.