THE POWER RANK

3 surprising college football teams by adjusted yards per play

By 4 Comments

You need an accurate assessment of college football teams. In particular, you need to know which teams have improved or declined in 2017 and adjust their rating appropriately.

Team rankings is a good start. The Power Rank’s college football team rankings take the margin of victory in games and adjust for strength of schedule through an algorithm I developed.

However, points based rankings can be misleading. Turnovers can leave their imprint on the final score, as we will see in discussing Wisconsin below. As much as you might think teams control turnovers, analytics suggests randomness plays a big role in these pivotal plays.

To get an alternative assessment of college football teams, we’ll look at a different efficiency metric.

College football rankings by adjusted yards per play

Below, I show my team rankings based on yards per play adjusted for strength of schedule. To calculate these rankings, I first rank each offense and defense based on yards per play, a powerful efficiency upon which turnovers have little impact. The ranking algorithm makes the adjustments for schedule.

The offense and defense rankings get combined into one team rank in the list below. Unless otherwise noted, all rankings referred to in this article are offense, defense or team rankings by these adjusted yards per play.

The rankings below reflect data from only the 2017 season. As teams only have played at most 7 games, this reflects a small sample size. In addition, yards per play is not immune from randomness. That 50 yard pass play that went through the hands of the defender counts.

However, these rankings do provide insight into the rise and fall of teams. Here are the top 25 teams. The number gives an expected net yards per play difference against an average FBS opponent.

1. Oklahoma, 3.26
2. Oklahoma State, 2.95
3. Alabama, 2.93
4. Ohio State, 2.90
5. Georgia, 2.27
6. Washington, 2.11
7. Wisconsin, 2.09
8. Auburn, 2.01
9. Florida State, 1.98
10. Stanford, 1.84
11. Michigan, 1.84
12. Miami (FL), 1.79
13. Texas Tech, 1.78
14. Clemson, 1.73
15. UCF, 1.72
16. Notre Dame, 1.45
17. Wake Forest, 1.38
18. Oregon, 1.36
19. Penn State, 1.34
20. Louisville, 1.32
21. USC, 1.28
22. Washington State, 1.28
23. Michigan State, 1.22
24. North Carolina State, 1.16
25. LSU, 1.14

While this top 25 list has many stories, two teams stand out to me.

Wisconsin

The Badgers seemed to struggle this past weekend against Purdue. They won 17-9, but needed a last second defensive stand to a win a game in which they were favored by 17.

However, the final score doesn’t reflect what happened in the game. Wisconsin gained 494 yards on 6.96 yards per play. In contrast, Purdue had 221 yards on 4.02 yards per play.

Turnovers played a big role on the scoreboard, as the Badgers gave away the ball 3 times. Freshman running back Jonathan Taylor fumbled at the Purdue 4 yard line late in the game.

Wisconsin started the season 8th in my preseason rankings based on points. They have dropped 18th after 7 weeks of the season.

However, Wisconsin ranks 7th as a team by adjusted yards per play. The defense has been excellent as usual, but the gains come from the offense.

Wisconsin ranked 54th on offense in 2016 but have jumped to 9th over the first 7 weeks of 2017. Taylor has led a resurgent rush attack by gaining 7.8 yards per carry.

UCF

I laughed when I read Bill Connelly’s preview of UCF this preseason. He argued that everything head coach Scott Frost touches turns to gold based on his track record.

Dude, did you look at UCF’s offensive numbers last season? They ranked 115th in my adjusted yards per play. They ranked 117th in your own S&P+ numbers!

Then the season started, and UCF ended up first in my points based rankings after 4 weeks. I explained why this was a fluke, and why they would fall in the coming weeks.

UCF has fallen in my points based rankings, but only to 7th. They continue to plow through soft AAC competition.

The team rankings by adjusted yards per play suggests their improvement is real, as they rank 15th through 7 weeks of 2017. The offense has gone from the bottom of FBS to 6th this season, all with the same quarterback in McKenzie Milton.

The defense has slipped a bit. They ranked 19th last season in leading UCF to an unexpected bowl berth but has slipped to 49th this season.

Still, Bill is having the last laugh about Scott Frost.

Members of The Power Rank have access to these yards per play rankings broken down into offense and defense. You also get predictions for spreads and totals in college football and the NFL.

To learn more, click on the image or here.

The next 25 college football teams

Let’s look at the next 25 teams in my team rankings by adjusted yards per play.

26. TCU, 1.07
27. North Texas, 1.02
28. Mississippi State, 0.90
29. Utah, 0.79
30. Boise State, 0.69
31. Georgia Tech, 0.68
32. Mississippi, 0.68
33. Fresno State, 0.67
34. San Diego State, 0.66
35. Appalachian State, 0.65
36. Iowa State, 0.64
37. Arizona, 0.62
38. Purdue, 0.60
39. Texas, 0.47
40. Nebraska, 0.43
41. Indiana, 0.43
42. Kansas State, 0.41
43. UTSA, 0.41
44. Virginia Tech, 0.38
45. California, 0.37
46. UCLA, 0.37
47. Northern Illinois, 0.32
48. Southern Miss, 0.31
49. Minnesota, 0.31
50. UNLV, 0.31

This leads us to one of the best stories in college football this season.

Fresno State

How the hell is Fresno State 33rd?

The Bulldogs struggled through an awful 1-11 season in 2016, as coach Tim DeRuyter got fired during the season. They brought in former Cal coach Jeff Tedford to save the program.

Tedford had some brilliant offenses at Cal. However, it didn’t look he could do too much with a unit that ranked dead last in the FBS in 2016.

Fresno State started the season 1-2, with two losses to playoff contenders Alabama and Washington. Then Tedford turned to QB Marcus McMaryion, a player who transferred from Oregon State in August.

The offense has been stellar in winning 3 games since the change. Fresno State has gained 7.1 and 8.9 yards per play against Nevada and New Mexico respectively. The offense ranks 23rd based on data in 2017.

The offensive coordinator also plays a role in this story. Tedford brought in Kalen DeBoer, who had worked miracles in getting Eastern Michigan’s offense to rank 67th in 2016.

Eastern Michigan’s offense has fallen off a cliff in 2017 (120th), while Fresno State has surged.

Can Fresno State keep it up? Remember, these rankings are based on a small sample size, and Fresno State thumped some bad defenses in Nevada and New Mexico. You have reason to doubt whether a QB that couldn’t win the job at Oregon State can keep Fresno State as a top 25 offense.

These are the rest of the team rankings by adjusted yards per play.

51. West Virginia, 0.30
52. Florida, 0.26
53. Memphis, 0.22
54. South Carolina, 0.19
55. New Mexico State, 0.18
56. Maryland, 0.15
57. Colorado State, 0.14
58. Northwestern, 0.14
59. Buffalo, 0.14
60. Iowa, 0.12
61. Missouri, 0.10
62. Middle Tennessee State, 0.10
63. Houston, 0.09
64. Arizona State, 0.08
65. Texas A&M, 0.04
66. Florida Atlantic, 0.04
67. SMU, 0.03
68. Syracuse, 0.02
69. Virginia, 0.01
70. Troy, -0.08
71. New Mexico, -0.11
72. Idaho, -0.14
73. Ohio, -0.19
74. Duke, -0.21
75. South Florida, -0.22
76. Colorado, -0.24
77. Navy, -0.31
78. Marshall, -0.35
79. North Carolina, -0.36
80. Arkansas, -0.37
81. Tennessee, -0.38
82. Vanderbilt, -0.42
83. Boston College, -0.53
84. Baylor, -0.55
85. Oregon State, -0.57
86. Eastern Michigan, -0.58
87. Toledo, -0.61
88. Kentucky, -0.65
89. UAB, -0.69
90. Utah State, -0.72
91. Florida International, -0.81
92. Rutgers, -0.82
93. Louisiana Tech, -0.82
94. Air Force, -0.85
95. Arkansas State, -0.85
96. Brigham Young, -0.89
97. South Alabama, -0.90
98. Western Michigan, -0.94
99. Pittsburgh, -0.97
100. Connecticut, -0.97
101. Wyoming, -0.99
102. Hawaii, -1.02
103. Miami (OH), -1.04
104. Nevada, -1.06
105. Cincinnati, -1.11
106. Massachusetts, -1.14
107. Kansas, -1.15
108. Tulane, -1.16
109. Coastal Carolina, -1.17
110. Central Michigan, -1.19
111. Temple, -1.22
112. San Jose State, -1.25
113. Tulsa, -1.33
114. Illinois, -1.37
115. Bowling Green, -1.40
116. Army, -1.46
117. Louisiana Monroe, -1.51
118. Akron, -1.56
119. Georgia State, -1.67
120. Old Dominion, -1.68
121. Georgia Southern, -1.69
122. Louisiana Lafayette, -1.70
123. Charlotte, -1.83
124. Western Kentucky, -1.93
125. Kent State, -1.99
126. Texas State, -2.08
127. Rice, -2.09
128. East Carolina, -2.14
129. Ball State, -2.27
130. UTEP, -2.40

The problem with using only one model in football predictions

By Leave a Comment

You might wonder why UCF is first in my public college football rankings after four weeks of the 2017 season. That makes no sense whatsoever.

Let’s break it down. First, UCF has played only two games this season due to the hurricanes in Florida. They stomped Florida International 61-17, and then they beat Maryland on the road 38-10, a game in which they were a 4 point underdog.

My team ranking algorithm takes margin of victory and adjusts for strength of schedule. UCF not only has a huge margin of victory, but they have a key win over Maryland.

Maryland makes UCF look good because they beat Texas 51-41 in their opener. The team rankings don’t know that Maryland QB Tyrell Pigrome got hurt in the Texas game and won’t play the remainder of the season.

Texas went to USC and almost pulled off the upset in week 3. This boosts the stature of Texas, which gives Maryland a bump and pushes UCF up higher in my team rankings.

USC destroyed a good Stanford team, which pushes USC up in my rankings, etc. This propagates all the way to UCF.

Using only games this year, UCF is rated 56.8 points better than FBS average, an unsustainable level of play even for Alabama.

Now, these team calculations get blended with other points based numbers from previous years to give you the public rankings on this site. A regression model determines the parameters.

Usually, the parameters work pretty well. Most of the remaining teams make sense. However, it completely fails for UCF. They benefit from playing only two games, and they get a big strength of schedule boost from the Maryland win.

Here’s the take home message: it’s not good enough to use just points based metrics to make football predictions.

Members of The Power Rank have access to ensemble predictions that combine these points based numbers with other calculations based on yards per play and closing market spreads. UCF has a much more reasonable rank of 47th in these numbers. The resulting prediction against Memphis also makes sense.

In addition, members have access to totals predictions for both college football and the NFL for this weekend, the first of the season.

To become a member of The Power Rank, click here.

Podcast: Drew Martin on Predicting College Football for Small Schools

By Leave a Comment

On this episode of The Football Analytics Show, I welcome Drew Martin from Sports Book Review. He’s also my co-host on the new show Man vs Machine, where we contrast his football based approach to handicapping with my numbers.

Among other topics, we discuss:

  • Why he decided to leave JP Morgan to enter the sports world
  • How being a first team all county QB at a Florida high school helps his handicapping
  • His process for studying teams from the Sun Belt, Conference USA and the MAC
  • How the show Man vs Machine originated
  • The Florida team Drew thinks will have value in the markets

After the interview, I do my own segment on small sample size, and how it might affect the NFL games you want to watch in Week 2. Listen at 22:30.

To listen on iTunes, click here.

To listen here, click on the right pointing triangle:

The reason you can’t avoid the curse of small sample size

By Leave a Comment

Cincinnati opened the 2017 season with Austin Peay, a FCS team that went in 0-11 the previous year. It didn’t result in the expected blow out, as Cincinnati only won 26-14.

In contrast, Michigan looked good in their opener against Florida. It didn’t seem like Cincinnati had a chance in traveling to the Big House for week 2.

Instead, Cincinnati made it a game, as they trailed only 17-14 late in the third quarter before Michigan pulled away.

The Texas defense got gashed in their first game, allowing upstart Maryland to gain 8.3 yards per play. It seemed like miracle worker turned Texas coach Tom Herman had lost his touch in a 51-41 loss.

However, the Texas defense shut out San Jose State in week 2. It probably shouldn’t be a surprise, given that the defense ranked 34th in my adjusted yards per play in 2016 and brought back lots of talent.

The curse of small sample size. You know you shouldn’t make a quick judgment, even if you watched every play.

But somehow we all have problems with this.

In my article The Reason You Can’t Avoid The Curse of Small Sample Size, I discuss the human psychology behind why we have such a difficult relationship with randomness. It involves the very trait that makes us most human.

To check out the article, click here.

It also have some hilarious anecdotes from my volunteer teaching at the Summers-Knoll School in Ann Arbor.

Podcast: Bill Connelly on College Football Predictions for Week 2, 2017

By Leave a Comment

It might as well be New Year’s Day with this week’s college football games.

  • Oklahoma at Ohio State
  • Stanford at USC
  • Auburn at Clemson
  • Georgia at Notre Dame

On the latest episode of The Football Analytics Show, I discuss the predictions for these games with Bill Connelly, college football analytics expert and SB Nation writer extraordinaire.

In addition to breaking down these games, we also discuss how he has written a preview for each FBS team for seven years running. Bill also describes his research on the importance of success rate in predicting football games.

After the interview with Bill, I come back with a short segment on the Oakland Raiders, a team that I think the markets have wrong. Listen at 36:00.

To listen The Football Analytics Show on iTunes, click here.

To listen here, click on the right pointing triangle.