Are yardage and points really the best way to determine a defense's (or offense's) performance? I've come to be a firm believer in that it's not these figures that determines a unit's worth- but stops. In the era of up-tempo, offenses should naturally be expected to score more because they get the ball more and defensive statistics should deteriorate as well because they are having to defend more possessions. At the end of the day I think it's how many times you can get the ball back to your offense with limited damage done. Whether it be by forcing punt, recovering a fumble, or forcing a missed field goal; how ever many more drives you can steal for your offense is what makes a good defense. For instance, Auburn surrendered similar yardage in the 2013 and 2014 Iron Bowls. Huge difference in the outcome, though. And if a defense is given a short field, and they force a FG, should those points really be counted against them? Or should they actually be rewarded?
So I came up a "point system" of sorts that awards different points based drive-by-drive. For instance, if a defense forced a punt they are awarded a +1. A 3-n-out is worth extra, +1.25. They are higher values for turnovers, defensive touchdowns, and safeties. Then there are negative points awarded for touchdowns allowed, long field goal drives, and even more for lengthy TD drives. At the end, the total is divided by number of drives and you get your average score for the game. It's built so that a .500 is considered an "average" performance. Coupled with an average offensive performance, you have a 50-50 shot of winning. The closer you get to 1.00, the better the performance. Past 1.00 is a truly exceptional defensive effort. And the further towards 0.00, the worse you are.
To illustrate what I'm talking about, here's the drive chart from 2010 Auburn-LSU...
1. Interception: +2 points awarded
2. FG drive of >40 yards: -0.5 points
3. Punt: +1
4. TD drive of >75 yards: -1.5
5. 3-and-Out: +1.25
6. 3-and-Out: +1.25
7. Punt: +1
8. 3-and-Out: +1.25
9. TD drive of 30-74 yards: -1
10. 3-and-Out: +1.25
11. Turnover on Downs: +1.5
Total: 7.5 points / 11 drives: 0.68181
So paired with a relative offensive performance, the defense gives you a 68.18% probability of a W.
The same thing can be done for offense, just points are now awarded for scores and taken away for punts/turnovers. In this particular game, Auburn's offense had a scale score of .5555, or just above average. Combine this with the defensive performance, and Auburn gave themselves a win probability of 61.87% based on their offensive and defensive performance. This, of course, doesn't take into account some of the whacky events that can take place during a football game: head coaching decisions, special teams, officiating, etc. It's just a measure of how well a team plays on both sides of the ball. I actually did this for the entirety of the 2010 season and the results were pretty interesting.
OPPONENT: DEFENSIVE SCALE SCORE / OFFENSIVE SCALE SCORE / WIN PROBABILITY
Arkansas State: .5167 / .9423 / 72.95% (Projected 15 point win)
Mississippi State: 1.0417 / .1944 / 61.80% (+6)
Clemson: .4231 / .2692 / 34.62% (-8)
South Carolina: .5682 / .8611 / 71.47% (+14)
Louisiana-Monroe: 1.375 / 1.425 / 99.9% (+40)
Kentucky: .025 / 1.075 / 55% (+3)
Arkansas: .3462 / 1.3542 / 85.02% (+22)
LSU: .6818 / .5555 / 61.97% (+6)
Ole Miss: .2917 / 1.3 / 79.69% (+18)
Chattanooga: .3636 / 1.6136 / 99.7% (+39)
Georgia: (-).125 / 1.325 / 60% (+5)
Bama: .6042 / .5454 / 57.48% (+4)
South Carolina: .625 / 1.572 / 99.9% (+40)
Oregon: .8517 / .2292
So over the course of the entire season, the 2010 Auburn defense averaged .5422 a game and .4510 versus SEC opponents. This essentially says they were a slightly above average D-1 unit and slightly below average
in the conference. In contrast, the offense had a whopping .9441 score. Combined together, you get an overall TEAM SCORE of roughly .7432. This means that if 2010 Auburn played that regular season schedule 100 times, 9-3 would be the most common result. But we were really only actually outplayed once that year according to the system: Clemson. The defense carried the team twice against Mississippi St and Oregon. They outplayed the offense 5 times: Miss St, Clem, LSU, Bama, and Oregon.
For fun, let's compare the 2010 team with the 2013 Auburn Tigers;
Washington State: .7857 / .4545 / 62.01% (+6)
Arkansas State: .9545 / .8181 / 88.63% (+26)
Mississippi State: .4318 / .3125 / 37.23% (-6)
LSU: .4231 / (-).0333 / 19.49% (-18)
Ole Miss: .6786 / .3409 / 50.98% (+1)
W Carolina: 1.068 / 1.568 / 99.9% (+52)
Texas A&M: .1964 / .8846 / 54.05% (+2)
FAU: 1.058 / .9375 / 99.77% (+37)
Arkansas: .800 / .9722 / 88.61% (+26)
Tennessee: .7083 / .950 / 82.92% (+20)
Georgia: .1042 / 1.1818 / 64.30% (+7)
Bama: .3125 / .4167 / 36.46% (-7)
Mizzou: .2857 / 1.089 / 68.74% (+19)
FSU: .4375 / .5625 / 50% (EVEN)
TOTAL TEAM EFFICIENCY
2010: 74.32
2013: 66.79
DEFENSIVE SCORE SCALE:
2010: 54.22
2013: 58.89
DEFENSIVE SCORE VS. SEC OPPONENTS:
2010: 45.1
2013: 43.78
OFFENSIVE SCORE SCALE:
2010: 94.41
2013: 74.68
OFFENSIVE SCORE VS. SEC OPPONENTS:
2010: 97.1
2013: 67.94
PROJECTED RECORD:
2010: 8.92-3.08
2013: 8.01-3.99
I think it'll be interesting to see how these numbers compare to other past teams, such as the 2004 squad.
So I came up a "point system" of sorts that awards different points based drive-by-drive. For instance, if a defense forced a punt they are awarded a +1. A 3-n-out is worth extra, +1.25. They are higher values for turnovers, defensive touchdowns, and safeties. Then there are negative points awarded for touchdowns allowed, long field goal drives, and even more for lengthy TD drives. At the end, the total is divided by number of drives and you get your average score for the game. It's built so that a .500 is considered an "average" performance. Coupled with an average offensive performance, you have a 50-50 shot of winning. The closer you get to 1.00, the better the performance. Past 1.00 is a truly exceptional defensive effort. And the further towards 0.00, the worse you are.
To illustrate what I'm talking about, here's the drive chart from 2010 Auburn-LSU...
1. Interception: +2 points awarded
2. FG drive of >40 yards: -0.5 points
3. Punt: +1
4. TD drive of >75 yards: -1.5
5. 3-and-Out: +1.25
6. 3-and-Out: +1.25
7. Punt: +1
8. 3-and-Out: +1.25
9. TD drive of 30-74 yards: -1
10. 3-and-Out: +1.25
11. Turnover on Downs: +1.5
Total: 7.5 points / 11 drives: 0.68181
So paired with a relative offensive performance, the defense gives you a 68.18% probability of a W.
The same thing can be done for offense, just points are now awarded for scores and taken away for punts/turnovers. In this particular game, Auburn's offense had a scale score of .5555, or just above average. Combine this with the defensive performance, and Auburn gave themselves a win probability of 61.87% based on their offensive and defensive performance. This, of course, doesn't take into account some of the whacky events that can take place during a football game: head coaching decisions, special teams, officiating, etc. It's just a measure of how well a team plays on both sides of the ball. I actually did this for the entirety of the 2010 season and the results were pretty interesting.
OPPONENT: DEFENSIVE SCALE SCORE / OFFENSIVE SCALE SCORE / WIN PROBABILITY
Arkansas State: .5167 / .9423 / 72.95% (Projected 15 point win)
Mississippi State: 1.0417 / .1944 / 61.80% (+6)
Clemson: .4231 / .2692 / 34.62% (-8)
South Carolina: .5682 / .8611 / 71.47% (+14)
Louisiana-Monroe: 1.375 / 1.425 / 99.9% (+40)
Kentucky: .025 / 1.075 / 55% (+3)
Arkansas: .3462 / 1.3542 / 85.02% (+22)
LSU: .6818 / .5555 / 61.97% (+6)
Ole Miss: .2917 / 1.3 / 79.69% (+18)
Chattanooga: .3636 / 1.6136 / 99.7% (+39)
Georgia: (-).125 / 1.325 / 60% (+5)
Bama: .6042 / .5454 / 57.48% (+4)
South Carolina: .625 / 1.572 / 99.9% (+40)
Oregon: .8517 / .2292
So over the course of the entire season, the 2010 Auburn defense averaged .5422 a game and .4510 versus SEC opponents. This essentially says they were a slightly above average D-1 unit and slightly below average
in the conference. In contrast, the offense had a whopping .9441 score. Combined together, you get an overall TEAM SCORE of roughly .7432. This means that if 2010 Auburn played that regular season schedule 100 times, 9-3 would be the most common result. But we were really only actually outplayed once that year according to the system: Clemson. The defense carried the team twice against Mississippi St and Oregon. They outplayed the offense 5 times: Miss St, Clem, LSU, Bama, and Oregon.
For fun, let's compare the 2010 team with the 2013 Auburn Tigers;
Washington State: .7857 / .4545 / 62.01% (+6)
Arkansas State: .9545 / .8181 / 88.63% (+26)
Mississippi State: .4318 / .3125 / 37.23% (-6)
LSU: .4231 / (-).0333 / 19.49% (-18)
Ole Miss: .6786 / .3409 / 50.98% (+1)
W Carolina: 1.068 / 1.568 / 99.9% (+52)
Texas A&M: .1964 / .8846 / 54.05% (+2)
FAU: 1.058 / .9375 / 99.77% (+37)
Arkansas: .800 / .9722 / 88.61% (+26)
Tennessee: .7083 / .950 / 82.92% (+20)
Georgia: .1042 / 1.1818 / 64.30% (+7)
Bama: .3125 / .4167 / 36.46% (-7)
Mizzou: .2857 / 1.089 / 68.74% (+19)
FSU: .4375 / .5625 / 50% (EVEN)
TOTAL TEAM EFFICIENCY
2010: 74.32
2013: 66.79
DEFENSIVE SCORE SCALE:
2010: 54.22
2013: 58.89
DEFENSIVE SCORE VS. SEC OPPONENTS:
2010: 45.1
2013: 43.78
OFFENSIVE SCORE SCALE:
2010: 94.41
2013: 74.68
OFFENSIVE SCORE VS. SEC OPPONENTS:
2010: 97.1
2013: 67.94
PROJECTED RECORD:
2010: 8.92-3.08
2013: 8.01-3.99
I think it'll be interesting to see how these numbers compare to other past teams, such as the 2004 squad.
