A new way to study offensive/defensive statistics
Published: Monday, 14 January 2013 12:34
by Zach Barnett
Earlier this month we praised the Buffalo Bills for bringing analytics to the NFL. As we said then, the study of advanced statistics have found their homes in baseball and basketball, and it was long past due that they find a place in football as well.
Like almost everything else in football, the college game has beaten the NFL to the punch. Over the weekend we received a gold mine of data from Tulsa graduate assistant Zak Bigelow (who was the FootballScoop GA of the Day back in June) has developed a new way to evaluate the efficiency of offenses and defenses in football. For anyone familiar with the Points Per 100 Possessions metric in basketball, Bigelow created football's answer: points per 75 plays.
Bigelow's intent was to create a set of data that, "does not give offenses points for defensive or special teams touchdowns and punishes them for allowing touchdowns off of interceptions and fumbles," while also accounting for pace. Bigelow began this study following the 2011 season and settled 75 plays because the average number of snaps in the 2011 season was 144.
Before we delve into the numbers, understand that, for example, Baylor ran 1,072 plays in 13 games, which put their defense on the field for 1,080 plays. Conversely, Army ran 921 plays in 12 games, but its triple-option offense only required the Black Knights' defense to be on the field for 763 plays. So, over the course of a season, Baylor was tasked with defending 317 more plays than Army.
Offensive Points Per 75 Plays
1. Oregon 45.68
2. Alabama 45.27
3. Kansas State 45.04
4. Louisiana Tech 43.98
5. Oklahoma State 43.93
Defensive Points Per 75 Plays
1. Alabama 13.71
2. Notre Dame 15.00
3. Utah State 15.50
4. Rutgers 15.65
5. Florida 16.45
It's always important to remember that advanced stats aren't to be used as a sole evaluation point, but as a reference point to what we see on the field. Anyone who watched college football knew the above teams had excellent offenses and defenses. The only eye-opener was just how efficient the Crimson Tide offense was.
If we look beyond the numbers, which units were actually more efficient than the numbers gave them credit for?
Offensive Points Per 75 Plays vs. Traditional Rankings
Central Michigan - jumped 27 spots. Ranked 64th nationally (28.8 ppg) in traditional rankings, 37th (33.97) in points/75 plays
Michigan - 26 spots. 57th (29.8 ppg) in traditional rankings, 31st (35.49) in points/75 plays
Texas State - 22 spots. 67th (28.6 ppg) in traditional rankings, 45th (33.24) in points/75 plays
Western Kentucky - 21 spots. 70th (28.2 ppg) in traditional rankings, 49th (32.64) in points/75 plays
Mississippi State - 21 spots. 60th (29.5 ppg) in traditional rankings, 39th (33.87) in points/75 plays
Here are some offenses that looked worse when judged on efficiency.
Marshall - dropped 31 spots. 7th (40.9 ppg) in traditional rankings, 38th (33.88) in points/75 plays
Houston - dropped 28 spots. 38th (32.4 ppg) in traditional rankings, 66th (29.65) in points/75 plays
Tulsa - dropped 27 spots. 30th (34.7 ppg) in traditional rankings, 57th (31.42) in points/75 plays
Ball State - dropped 23 spots. 33rd (33.6 ppg) in traditional rankings, 56th (31.67) in points/75 plays
Rice - dropped 21 spots. 41st (31.8 ppg) in traditional rankings, 62nd (30.26) in points/75 plays
Going back to the basketball comparison, the offenses in the below group were volume shooters. They put a lot of points on the scoreboard, but took a lot of reps to get there. The bottom five teams each ran 984 plays or more, while the top group all ran 848 plays or fewer.
Defensive Points Per 75 Plays vs. Traditional Rankings
Houston - jumped 34 spots. 110th nationally (36 ppg) in traditional rankings, 75th (31.43) in points/75 plays
Boston College - 25 spots. 75th (29.7 ppg) in traditional rankings, 50th (27.27) in points/75 plays
Arizona - 25 spots. 104th (35.3 ppg) in traditional rankings, 79th (31.73) in points/75 plays
Miami - 24 spots. 83rd (30.5 ppg) in traditional rankings, 59th (28.74) in points/75 plays
California - 17 spots. 98th (33.1 ppg) in traditional rankings, 81st (31.84) in points/75 plays
Baylor - 17 spots. 113th (37.2 ppg) in traditional rankings, 96th (33.61) in points/75 plays
Conversely, here are some defenses that looked much worse when examined on a per play basis
New Mexico - dropped 24 spots. 79th (30.2 ppg) in traditional rankings, 103rd (34.88) in points/75 plays
Florida Atlantic - dropped 20 spots. 85th (30.8 ppg) in traditional rankings, 105th (35.13) in points/75 plays
Georgia Tech - dropped 20 spots. 65th (28.3 ppg) in traditional rankings, 85th (32.28) in points/75 plays
Louisville - dropped 20 spots. 36th (23.8 ppg) in traditional rankings, 56th (27.89) in points/75 plays
Army - dropped 16 spots. 104th (35.3 ppg) in traditional rankings, 120th (41.68) in points/75 plays
Again, the data is backing up logical sense. Defenses that line up across explosive up-tempo offenses week after week (Houston, Arizona, California, Baylor) look better than the traditional stats say they are, while defenses that are aided by clock-chewing triple-option offenses (New Mexico, Georgia Tech, Army) look worse.
The only team to show up in both studies, Houston turns out to be the perfect test case in the points per 75 plays study. The Cougars' offense wasn't as good as the numbers dictate, but its defense was better than the numbers say. In the end, though, they add up to what Houston ended up becoming, a 5-7 team.
When studying not just the numbers themselves, but how they stack up when compared to the traditional statistics, only the middle class tended to be greatly effected. Great offenses and stingy defenses look good under any microscope. But points per 75 plays succeeds at what it intends to accomplish, stripping away all the clutter, in this case, pace, special teams touchdowns, etc., and revealing who really makes the most of their time spent on the field.