Best defenses in College Football

[sciencewhiz]

It was mentioned a few weeks ago that it's easy to have a good defense statistically, if you haven't faced any good offenses. A better way to determine the best defenses is to look at their opponents average yardage and compare that to how many yards the defense gave up.

For example, The best offense Alabama has faced is Arkansas who is 55th in the nation. USC has faced 5 offenses in the top 20.

The top defenses by difference from opponents average offensive yardage:
#
Rank Team (defensive yardage rank), difference
1. Ohio St (7), -156.2
2. Miami (1), -124.4
3. Penn St (14), -111.5
4. Virginia Tech (2) -106.0
5. Texas (6), -96.9
6. Michigan (51), -93.3
7. Tennessee (17), -86.5
8. Alabama (3), -75.8
9. USC (41), -74.2
10. Oklahoma (15), -72.2

 

[MikeMike]

I think Penn State vs. ND would be a good game, but i honestly dont think ND has a chance on the offensive side of the ball. Penn State has a smothering D.

same goes for Ohio State vs. ND.

 

[rocadelpunk]

go bucks.

 

[sao123]

we always expect psu to have a top rated defense....
hence the term "linebacker u"

 

[dullard]

Nice ranking there. I am the type of person who argues that statistically you can have a great offense/defense by simply playing opponents with a bad defense/offense.

One important question though, is how do you statistically measure a good offense? In your list, you used yards/game as the measurement of a good offense. While yardage is certainly important, that one statistic gives many bad teams a good rating. Teams like Houston and Hawaii are horrible teams, yet have one of the highest yards/game in the nation. And teams that really are decent (not great but decent) are at the bottom of that list: Oklahoma, Kansas St, Tennessee, etc.

So yardage alone isn't enough. Points also matter. In fact, people could argue that points are all that matters. So you could do that same list that you just did, but use points instead of yards.

And then there are the odd combinations:
[*]A ranking of (Points/game)*(Yards/game) covers both important aspects. In fact, a simple sorted list of that statistic is remarkably close to the computer rankings of teams - ignoring all defensive or win/loss statistics. I do this on occasion just for fun. No, the match isn't perfect, but it is an interesting ranking nonetheless. Top of that list: USC, Texas Tech, Texas.
[*]My own favorite ranking is what I call the "get it done" statistic. Certain teams just happen to get the job done and score points when they aren't playing well. That is what often wins national championships. How do I measure this? Points/yard. These teams just happen to win in situations when they shouldn't. Top of that list? Fresno St and Texas.
[*]Points^2/yard. Combination of the two above. These teams are the teams with high powered offenses that win games. Top of the list: Texas, USC, Texas Tech.

I don't have any statistical evidence of which method of measuring a good offense is the best method. But if I were a betting man, I'd bet on P^2/Y.

 

[Queasy]

Originally posted by: sciencewhiz
For example, The best offense Alabama has faced is Arkansas who is 55th in the nation. USC has faced 5 offenses in the top 20.

Conversely, Alabama has faced off against Top 25 defenses througout the season (with two more upcoming) while USC has yet to face a defense in the top 50 and probably won't in the final weeks.

 

[nycxandy]

Originally posted by: dullard
Nice ranking there. I am the type of person who argues that statistically you can have a great offense/defense by simply playing opponents with a bad defense/offense.

One important question though, is how do you statistically measure a good offense? In your list, you used yards/game as the measurement of a good offense. While yardage is certainly important, that one statistic gives many bad teams a good rating. Teams like Houston and Hawaii are horrible teams, yet have one of the highest yards/game in the nation. And teams that really are decent (not great but decent) are at the bottom of that list: Oklahoma, Kansas St, Tennessee, etc.

So yardage alone isn't enough. Points also matter. In fact, people could argue that points are all that matters. So you could do that same list that you just did, but use points instead of yards.

And then there are the odd combinations:
[*]A ranking of (Points/game)*(Yards/game) covers both important aspects. In fact, a simple sorted list of that statistic is remarkably close to the computer rankings of teams quite well - ignoring all defensive or win/loss statistics. I do this on occasion just for fun. No, the match isn't perfect, but it is an interesting ranking nonetheless. Top of that list: USC, Texas Tech, Texas.
[*]My own favorite ranking is what I call the "get it done" statistic. Certain teams just happen to get the job done and score points when they aren't playing well. That is what often wins national championships. How do I measure this? Points/yard. These teams just happen to win in situations when they shouldn't. Top of that list? Fresno St and Texas.
[*]Points^2/yard. Combination of the two above. These teams are the teams with high powered offenses that win games. Top of the list: Texas, USC, Texas Tech.

I don't have any statistical evidence of which method of measuring a good offense is the best method. But if I were a betting man, I'd bet on P^2/Y.

Nicely done and I agree that there isn't one solid system, which encompasses different stats and factors, that gauges how good an offense is.

Also, a good offense which keeps the ball for the majority of the game can affect a team's play on defense.

 

[GeneValgene]

not surprised to see OSU at the top...

i remember when i was watching the texas game and seeing AJ Hawk for the first time...OMG he is a beast

 

[shud]

I think it's pretty amazing that USC is still playing backups and players converted from offense on the defensive side and is still ranked #9.

 

[nycxandy]

Can I see the formula you used?

 

[Nitemare]

LOL, who are the retards that are voting that USC has a defense?

Just goes to show you the reliability of polls.

 

[davew0670]

For example, The best offense Alabama has faced is Arkansas who is 55th in the nation. USC has faced 5 offenses in the top 20.

The top defenses by difference from opponents average offensive yardage:



too bad offensive yardage doesnt count on the scoreboard. Try offensive scoring against defensive scoring and see what you get.

 

[ElFenix]

could rank by scoring differential. and remove garbage time scores.

 

[Nitemare]

Originally posted by: nycxandy
Can I see the formula you used?

When team = USC, insert AlabamaD+1;

 

[edro]

Originally posted by: rocadelpunk

go bucks.

w3rd!

 

[vrbaba]

Originally posted by: nourdmrolNMT1
I think Penn State vs. ND would be a good game, but i honestly dont think ND has a chance on the offensive side of the ball. Penn State has a smothering D.

same goes for Ohio State vs. ND.

whens the game? penn vs Nd? and OSU vs ND?

are they on national TV? i need to make surei watch those.

 

[Feldenak]

Originally posted by: rocadelpunk


go bucks.

 

[sciencewhiz]

Originally posted by: Queasy

Originally posted by: sciencewhiz
For example, The best offense Alabama has faced is Arkansas who is 55th in the nation. USC has faced 5 offenses in the top 20.

Conversely, Alabama has faced off against Top 25 defenses througout the season (with two more upcoming) while USC has yet to face a defense in the top 50 and probably won't in the final weeks.

Oregon's just barely in the top 50 Fresno State is around 20th, and will be the best defense by far that USC will face until the bowls.

 

[sciencewhiz]

Originally posted by: nycxandy
Can I see the formula you used?

Pick a team (Ohio State for this example). Get a list of each team's opponents (Miami OH, Texas, San Diego St., Iowa, Penn St, Michigan St, Indiana, Minnesota, Illinois). For each opponent, find that team's total offense average and subtract how much they got against Ohio State. Using Texas, they average 518 yards but got 382 against Ohio State, giving a 132 yard difference. Do that for each game and average the results.

You can do the same thing with scoring, but as Elfenix alluded too, garbage time makes a much bigger difference in scoring statistics then yardage statistics.

 

[ElFenix]

Originally posted by: vrbaba

Originally posted by: nourdmrolNMT1
I think Penn State vs. ND would be a good game, but i honestly dont think ND has a chance on the offensive side of the ball. Penn State has a smothering D.

same goes for Ohio State vs. ND.

whens the game? penn vs Nd? and OSU vs ND?

are they on national TV? i need to make surei watch those.

i think they're speculating about a potential fiesta bowl match

 

[casvolsmu]

Originally posted by: Queasy

Originally posted by: sciencewhiz
For example, The best offense Alabama has faced is Arkansas who is 55th in the nation. USC has faced 5 offenses in the top 20.

Conversely, Alabama has faced off against Top 25 defenses througout the season (with two more upcoming) while USC has yet to face a defense in the top 50 and probably won't in the final weeks.

the problem there is this--the offenses in the SEC suck. It's not that the D's are good. After Alabama ran up a whopping 6-3 win against Tennessee, I didn't see ND having too tough a time moving the ball against UT. These stat arguments can always be skewed. If you play enough sh!tty offenses, your D stats will look great. Conversely, if your offense is good, the other teams will prepare differently, trying to actually score every possession, knowing it will take a TON of points to beat you. Play Alabama---10 points may be enough. No need for trickery or gutsy play-calling. Play Texas or USC--you better score some damn points. And often.

 

[bubbadu]

Originally posted by: Feldenak

Originally posted by: rocadelpunk


go bucks.

 

[cmdavid]

go canes!! our d is no-name d.... we dont have any stars, just a great job by our d-coordinator, randy shannon, and our team buying into his concepts...