[Goose]

While there is no such thing as a single number that can describe something as complex as how good an offense is, but I feel that the best attempt at such a thing is probably FEI, which is basically points per drive adjusted in complicated ways for strength of schedule and particularly bad things like turnovers.

While I can't claim to really "understand" FEI, I agree that what it attempts to do is probably the best "one number" approach I am aware of.

I worry about arguments concerning out/under-performing talent because there is too much of a chicken/egg issue there.

I agree, especially given the large number of NFL draft choices that don't work out. In the NFL you have full time professionals looking at players that are no longer "kids" playing on a somewhat level playing field (unlike High School, where the competition is often very unbalanced) and these players still out/under-perform regularly.

Another comment above was Stanford playing a bend but don't break defense. At least for our best defensive years of this stretch, that wasn't the case. Stats that would indicate more of an attacking, don't bend, defense would be a low yards/play and lot of sacks and tackles for loss.

As one might expect, here I disagree. I perhaps should have described Stanford's defensive strategy as read-react with special attention to preventing big plays (bend but don't break). Both read-react defenses and attacking defenses can turn in good stats like low yards/play and TFL. They use different methodology to accomplish it. Attacking defenses have a strategy they are going to use on a given play. It may be a blitz, it may be a stunt/stem, it may be base defense, but it is predetermined. A read/react defense may have a "plan" but depending on formation, personnel, tendencies, "keys" or gut feel, the plan may be changed before the snap. The defense will then "read" what the offense appears to be doing to choose their defensive response. 

That was very much Stanford in 2012. Stanford had Ben Gardner who was going to win some of the battles with the OL and make a TFL no matter what defense Stanford played, but basically the three first string defensive linemen were hard to move. They clogged up the middle and kept the linebackers "clean". All 4 starting linebackers were pretty good. One of those linebackers was Skov, who repeatedly demonstrated he knew what the offense was going to do before they did it. His anticipation on the snap count was legendary, but there is a video (that I can't find right now) entitled "Skov beats Oregon to the spot". He knew where the back was going and went there before he could be blocked. TFL. He wasn't the biggest, or the fastest (actually he was considered a bit "slow"), but he was always in exactly the right place. Read and react. As Mora said about our 2012 defense:

"But probably the most important thing is the way they pursue the football."

Against Oregon in 2012, Mariota actually completed quite a few passes but they were short passes. We "bent". But no long scoring passes. We didn't "break". So yes, a well-executed read and react defense can put up great statistics. It can be the best choice if you have somebody like Skov playing linebacker. Quoting him on the Oregon game:

"I think they had opportunities. We just did a great job of tackling and minimizing those opportunities. Guys were flying around. When they had a six-yard gain, with a chance to make it a 15- or 20-yard gain, we closed those windows and minimized that opportunities."

So I think that kind of speaks for itself. We undoubtedly were playing a read and react with special emphasis on limiting big plays. We were however doing it very well.

But I don't think any of this matters to discussions of whether slow down was a good strategy or not. I don't see any particular logic that having a good defense, or an attacking defense, has anything to do with whether a slow-down strategy is a good idea. 

Depth of the defense might. I don't remember the 2011-13 defenses being particularly thin. Starting in 2014 depth issues along the DL have been an almost constant concern. But that would be just one of many issues.

Certainly you can have a good defense that is an attacking defense, and certainly you can have one that is read-react. IMHO depth and skill set of the defense does have a large effect on how important it is to keep them rested. In 2011 the fact that Skov was not 100% unquestionably hurt our defense. In 2012 he was at full speed and it made a big difference. Several opponents commented on this. In the years 2011-2014, we had David Parry at Defensive Tackle. As I recall, he played the vast majority of the snaps as the years went on. We had several defensive ends during those years that were very good and passable depth behind them. Not so much at DT. Even as early as 2012 this was a bit of an issue.

[Goose]

While there is no such thing as a single number that can describe something as complex as how good an offense is, but I feel that the best attempt at such a thing is probably FEI, which is basically points per drive adjusted in complicated ways for strength of schedule and particularly bad things like turnovers.

While I can't claim to really "understand" FEI, I agree that what it attempts to do is probably the best "one number" approach I am aware of.

I worry about arguments concerning out/under-performing talent because there is too much of a chicken/egg issue there.

I agree, especially given the large number of NFL draft choices that don't work out. In the NFL you have full time professionals looking at players that are no longer "kids" playing on a somewhat level playing field (unlike High School, where the competition is often very unbalanced) and these players still out/under-perform regularly.

Another comment above was Stanford playing a bend but don't break defense. At least for our best defensive years of this stretch, that wasn't the case. Stats that would indicate more of an attacking, don't bend, defense would be a low yards/play and lot of sacks and tackles for loss.

As one might expect, here I disagree. I perhaps should have described Stanford's defensive strategy as read-react with special attention to preventing big plays (bend but don't break). Both read-react defenses and attacking defenses can turn in good stats like low yards/play and TFL. They use different methodology to accomplish it. Attacking defenses have a strategy they are going to use on a given play. It may be a blitz, it may be a stunt/stem, it may be base defense, but it is predetermined. A read/react defense may have a "plan" but depending on formation, personnel, tendencies, "keys" or gut feel, the plan may be changed before the snap. The defense will then "read" what the offense appears to be doing to choose their defensive response. 

That was very much Stanford in 2012. Stanford had Ben Gardner who was going to win some of the battles with the OL and make a TFL no matter what defense Stanford played, but basically the three first string defensive linemen were hard to move. They clogged up the middle and kept the linebackers "clean". All 4 starting linebackers were pretty good. One of those linebackers was Skov, who repeatedly demonstrated he knew what the offense was going to do before they did it. His anticipation on the snap count was legendary, but there is a video (that I can't find right now) entitled "Skov beats Oregon to the spot". He knew where the back was going and went there before he could be blocked. TFL. He wasn't the biggest, or the fastest (actually he was considered a bit "slow"), but he was always in exactly the right place. Read and react. As Mora said about our 2012 defense:

"But probably the most important thing is the way they pursue the football."

Against Oregon in 2012, Mariota actually completed quite a few passes but they were short passes. We "bent". But no long scoring passes. We didn't "break". So yes, a well-executed read and react defense can put up great statistics. It can be the best choice if you have somebody like Skov playing linebacker. Quoting him on the Oregon game:

"I think they had opportunities. We just did a great job of tackling and minimizing those opportunities. Guys were flying around. When they had a six-yard gain, with a chance to make it a 15- or 20-yard gain, we closed those windows and minimized that opportunities."

So I think that kind of speaks for itself. We undoubtedly were playing a read and react with special emphasis on limiting big plays. We were however doing it very well.

But I don't think any of this matters to discussions of whether slow down was a good strategy or not. I don't see any particular logic that having a good defense, or an attacking defense, has anything to do with whether a slow-down strategy is a good idea. 

Depth of the defense might. I don't remember the 2011-13 defenses being particularly thin. Starting in 2014 depth issues along the DL have been an almost constant concern. But that would be just one of many issues.

Certainly you can have a good defense that is an attacking defense, and certainly you can have one that is read-react. IMHO depth and skill set of the defense does have a large effect on how important it is to keep them rested. In 2011 the fact that Skov was not 100% unquestionably hurt our defense. In 2012 he was at full speed and it made a big difference. Several opponents commented on this. In the years 2011-2014, we had David Parry at Defensive Tackle. As I recall, he played the vast majority of the snaps as the years went on. We had several defensive ends during those years that were very good and passable depth behind them. Not so much at DT. Even as early as 2012 this was a bit of an issue.