[TrumpCard]

DL has had insufficient numbers from the time Diron Took over until last year’s class.  He recruited only 2 in 2017 and 2018, 1 in 2019, and 2 again in 2020. Things could have worked out better if the DLs that did come panned out at an excellent rate. Instead, Pakola basically never played and Fox and Keneley proved too skinny to fill a 3-4 DE spot. To boot, RJ and DWP weren’t great players and never matched their recruiting profile. Plus the 2020 guys were lightly recruited. The result is ONE really good DL in a 4 year window.

[TrumpCard]

DL has had insufficient numbers from the time Diron Took over until last year’s class.  He recruited only 2 in 2017 and 2018, 1 in 2019, and 2 again in 2020. Things could have worked out better if the DLs that did come panned out at an excellent rate. Instead, Pakola basically never played and Fox and Keneley proved too skinny to fill a 3-4 DE spot. To boot, RJ and DWP weren’t great players and never matched their recruiting profile. Plus the 2020 guys were lightly recruited. The result is ONE really good DL in a 4 year window.

[SamuelMcF]

We couldn't because we didn't have enough DL to put out there that could play well enough to do so. Instead, we ended up in nickel. Injuries contributed, but when you are trying to play Tucker Fisk on the DL, you have a problem.
So, my belief is that we have not recruited "well enough" to have an average defensive line. I think we both would agree we haven't had a DL that gave us confidence for a while. I am not arguing we don't have all the deficiencies you cite in player development etc. What I am arguing is that our recruiting was also deficient, and that class ranking by it's nature is not universally predictive of on-field performance for the reasons listed above.

Certainly there are many reasons, and both are true IMO.

Also, funny you mention Tucker Fisk as if that's a bad thing. Fisk took most of Ryan Johnson's snaps towards the end of last season and PFF graded him as our best DL, ahead of Thomas Booker. Imagine if he had started playing DL 2 seasons ago, as many were begging for him to. The need for another DL in the pipeline was glaringly obvious with the departures of Williams and Swann (I could go on an entire rant about "proactive" vs "reactive" recruiting, as you alluded to in your post), and we already had another solid one-dimensional blocking TE anyway at the time in Scooter Harrington. Little marginal value add on O, significant marginal value add on D since you need a DL rotation.

Another reason nobody has mentioned yet? Peter Hansen and his complete inability to coach ILBs (see: the six-man hockey shift "rotation" where Bobby Okereke and Jordan Perez saw the same amount of snaps). Good god.

ANOTHER reason nobody has mentioned yet? While I agree with BC's assertion that a cohort of 8ish adequate DBs across classes should be fine, we didn't have that either. How quickly folks forget the disaster at the safety position for what feels like forever. It was Justin Reid and a gaggle of converted corners, offensive players, and walk-ons. Kendall Williamson was a CB as a frosh. Malik Antoine was a CB for 2 years. Dallas Lloyd. Kodi Whitfield. Noah Williams. JJ Parson. "The best DB coach in college football" couldn't do half his job.

---

DL has had insufficient numbers from the time Diron Took over until last year’s class.  He recruited only 2 in 2017 and 2018, 1 in 2019, and 2 again in 2020. Things could have worked out better if the DLs that did come panned out at an excellent rate. Instead, Pakola basically never played and Fox and Keneley proved too skinny to fill a 3-4 DE spot. To boot, RJ and DWP weren’t great players and never matched their recruiting profile. Plus the 2020 guys were lightly recruited. The result is ONE really good DL in a 4 year window.

Phillips and Keneley were the 2020 guys, both were not "lightly recruited," and IMO it's too early to tell on Phillips. (This year we'll see if he's got anything, because, well, he has no choice.) You're thinking of the 2021s Franklin and Buckey, where it's clearly WAY too early to tell. But that's nitpicking an otherwise overall valid point.

[SamuelMcF]

We couldn't because we didn't have enough DL to put out there that could play well enough to do so. Instead, we ended up in nickel. Injuries contributed, but when you are trying to play Tucker Fisk on the DL, you have a problem.
So, my belief is that we have not recruited "well enough" to have an average defensive line. I think we both would agree we haven't had a DL that gave us confidence for a while. I am not arguing we don't have all the deficiencies you cite in player development etc. What I am arguing is that our recruiting was also deficient, and that class ranking by it's nature is not universally predictive of on-field performance for the reasons listed above.

Certainly there are many reasons, and both are true IMO.

Also, funny you mention Tucker Fisk as if that's a bad thing. Fisk took most of Ryan Johnson's snaps towards the end of last season and PFF graded him as our best DL, ahead of Thomas Booker. Imagine if he had started playing DL 2 seasons ago, as many were begging for him to. The need for another DL in the pipeline was glaringly obvious with the departures of Williams and Swann (I could go on an entire rant about "proactive" vs "reactive" recruiting, as you alluded to in your post), and we already had another solid one-dimensional blocking TE anyway at the time in Scooter Harrington. Little marginal value add on O, significant marginal value add on D since you need a DL rotation.

Another reason nobody has mentioned yet? Peter Hansen and his complete inability to coach ILBs (see: the six-man hockey shift "rotation" where Bobby Okereke and Jordan Perez saw the same amount of snaps). Good god.

ANOTHER reason nobody has mentioned yet? While I agree with BC's assertion that a cohort of 8ish adequate DBs across classes should be fine, we didn't have that either. How quickly folks forget the disaster at the safety position for what feels like forever. It was Justin Reid and a gaggle of converted corners, offensive players, and walk-ons. Kendall Williamson was a CB as a frosh. Malik Antoine was a CB for 2 years. Dallas Lloyd. Kodi Whitfield. Noah Williams. JJ Parson. "The best DB coach in college football" couldn't do half his job.

---

DL has had insufficient numbers from the time Diron Took over until last year’s class.  He recruited only 2 in 2017 and 2018, 1 in 2019, and 2 again in 2020. Things could have worked out better if the DLs that did come panned out at an excellent rate. Instead, Pakola basically never played and Fox and Keneley proved too skinny to fill a 3-4 DE spot. To boot, RJ and DWP weren’t great players and never matched their recruiting profile. Plus the 2020 guys were lightly recruited. The result is ONE really good DL in a 4 year window.

Phillips and Keneley were the 2020 guys, both were not "lightly recruited," and IMO it's too early to tell on Phillips. (This year we'll see if he's got anything, because, well, he has no choice.) You're thinking of the 2021s Franklin and Buckey, where it's clearly WAY too early to tell. But that's nitpicking an otherwise overall valid point.

[TrumpCard]

I got confused. I was thinking of 2021. So a 5 year period with just under 2 DLs per year on average. Two rated well below the top 1000, several more Edge than DL, and only ONE in 5 years who’s been even an average Pac-12 starter so far. I hold out hope for Phillips and Franklin though.

[TrumpCard]

I got confused. I was thinking of 2021. So a 5 year period with just under 2 DLs per year on average. Two rated well below the top 1000, several more Edge than DL, and only ONE in 5 years who’s been even an average Pac-12 starter so far. I hold out hope for Phillips and Franklin though.

[msqueri]

Thanks Goose. Thoughtful stuff, but we do disagree.

My analysis does not assume that class ranking should be absolutely predictive of expected performance for at least two reasons.

First, my observations of Stanford having at least average Pac-12 DL talent throughout Shaw's tenure are not based solely on recruiting rankings but also on the ultimate output of NFL players. We continue to send DL to the NFL, we just don't have line play at the college level that reflects the talent on the roster. Even Tucker Fisk, who you lament us having to play (and in any case wasn't a DL starter for us) is on an NFL roster. There's nothing below average about him as a third or fourth DL. Thomas Booker being a 4-star recruit, being all-conference twice, and then being drafted to the NFL are three different evaluations at four different points in time that all point to him being above-average for a number one DL on a Pac-12 team. I get the sense you think Dalyn Wade-Perry and Ryan Johnson had below-average talent and/or fit for second or third DL on a Pac-12 team, but there isn't much evidence for that either. They were 4-star recruits and, in Johnson's case, transferred elsewhere at the Power Five scholarship level. Whether you look at recruiting rankings or NFL evaluations, evidence does not point to Stanford having below-average DL talent. The one thing that arguably does is performance at Stanford. If that doesn't speak to coaching and scheme I don't know what does. 

Second, my observations on this point are not based on aggregate recruiting rankings, so many of your points about averaging across recruits/lower-ranked guys being obscured by higher-ranked guys, variance (why would Stanford have higher DL performance variance than other schools?), etc. are moot. Rather, I'm observing based on the individual guys we have. To use your frame, the weak links. Look at the non-Booker guys we have to play. Wade-Perry and Johnson were 4-star recruits and the last couple of years were extremely veteran former 4-star recruits. Nothing Plan B or Plan C about that.

In fact, this argument you're making about Stanford resorting to Plan B or Plan C DL does not reflect the reality of our roster ever under Shaw. Let's look at the DL in the years in which Anderson's defenses dropped off. The DL who have played 200 snaps for us between 2017-2021 have been Harrison Phillips, Dylan Jackson, Eric Cotton, Mike Williams, Jovan Swann, Thomas Booker, Dalyn Wade-Perry, Thomas Schaffer, Ryan Johnson, and Tucker Fisk. Four of the ten were 4-star recruits and the rest were high 3-stars. I don't spot a single Plan B or Plan C in the bunch. Really not sure where that trope is coming from. The closest you can get is Cotton and Fisk as position switches but they were coveted recruits whose future utility at DL was obvious at the time of initial recruitment.

Trying to give the weakest link argument some credence, the guys who have torpedoed our line play most egregiously have been Johnson, Jackson, and Cotton. Johnson had offers from Alabama, Auburn, Texas, Clemson, Georgia, Notre Dame, Oklahoma, LSU, Michigan - everybody. Jackson had offers from Alabama, LSU, Miami, Tennessee, Georgia Tech, etc. Cotton had offers from Washington State, Missouri, and the service academies. Not a single one was a Power Five reach or a Plan B and Johnson and Jackson were a level of recruit the majority of the Pac-12 would welcome as DL1, to say nothing of the DL3 we relied on them to be.

I think those who argue that something is below-average need to think systematically about what "average" means. All the arguments about weak links, variance, difficulty recruiting depth that allows DL to rotate, etc. apply to the majority of the Pac-12 as well. There is no way you can slice it - other than actual performance in college - that supports the idea Stanford has had below-average DL talent at any point under Shaw. Half the conference would trade for our DL room at the time of recruitment and NFL evaluators would pick Stanford's DL over half the conference at the NFL combine stage. That's really the bottom line.

While brevity is not my thing, if I had to summarize my view of our DL situation it would go like this: without talent like Solomon Thomas/Aziz Shittu and without coaching like Randy Hart, we fell off a cliff; the recruiting drop-off should not be ignored, but it's about not getting the difference makers (i.e., DL1) rather than the volume/variance arguments you're making (i.e., DL3/DL4). If you look at other Pac-12 teams the "average" DL3/DL4 situation is no prettier than ours; regardless of how much stock one puts in the recruiting, the evidence is overwhelming that our defense massively under-performs its talent, whether you look at recruiting or NFL evaluations.

[msqueri]

Thanks Goose. Thoughtful stuff, but we do disagree.

My analysis does not assume that class ranking should be absolutely predictive of expected performance for at least two reasons.

First, my observations of Stanford having at least average Pac-12 DL talent throughout Shaw's tenure are not based solely on recruiting rankings but also on the ultimate output of NFL players. We continue to send DL to the NFL, we just don't have line play at the college level that reflects the talent on the roster. Even Tucker Fisk, who you lament us having to play (and in any case wasn't a DL starter for us) is on an NFL roster. There's nothing below average about him as a third or fourth DL. Thomas Booker being a 4-star recruit, being all-conference twice, and then being drafted to the NFL are three different evaluations at four different points in time that all point to him being above-average for a number one DL on a Pac-12 team. I get the sense you think Dalyn Wade-Perry and Ryan Johnson had below-average talent and/or fit for second or third DL on a Pac-12 team, but there isn't much evidence for that either. They were 4-star recruits and, in Johnson's case, transferred elsewhere at the Power Five scholarship level. Whether you look at recruiting rankings or NFL evaluations, evidence does not point to Stanford having below-average DL talent. The one thing that arguably does is performance at Stanford. If that doesn't speak to coaching and scheme I don't know what does. 

Second, my observations on this point are not based on aggregate recruiting rankings, so many of your points about averaging across recruits/lower-ranked guys being obscured by higher-ranked guys, variance (why would Stanford have higher DL performance variance than other schools?), etc. are moot. Rather, I'm observing based on the individual guys we have. To use your frame, the weak links. Look at the non-Booker guys we have to play. Wade-Perry and Johnson were 4-star recruits and the last couple of years were extremely veteran former 4-star recruits. Nothing Plan B or Plan C about that.

In fact, this argument you're making about Stanford resorting to Plan B or Plan C DL does not reflect the reality of our roster ever under Shaw. Let's look at the DL in the years in which Anderson's defenses dropped off. The DL who have played 200 snaps for us between 2017-2021 have been Harrison Phillips, Dylan Jackson, Eric Cotton, Mike Williams, Jovan Swann, Thomas Booker, Dalyn Wade-Perry, Thomas Schaffer, Ryan Johnson, and Tucker Fisk. Four of the ten were 4-star recruits and the rest were high 3-stars. I don't spot a single Plan B or Plan C in the bunch. Really not sure where that trope is coming from. The closest you can get is Cotton and Fisk as position switches but they were coveted recruits whose future utility at DL was obvious at the time of initial recruitment.

Trying to give the weakest link argument some credence, the guys who have torpedoed our line play most egregiously have been Johnson, Jackson, and Cotton. Johnson had offers from Alabama, Auburn, Texas, Clemson, Georgia, Notre Dame, Oklahoma, LSU, Michigan - everybody. Jackson had offers from Alabama, LSU, Miami, Tennessee, Georgia Tech, etc. Cotton had offers from Washington State, Missouri, and the service academies. Not a single one was a Power Five reach or a Plan B and Johnson and Jackson were a level of recruit the majority of the Pac-12 would welcome as DL1, to say nothing of the DL3 we relied on them to be.

I think those who argue that something is below-average need to think systematically about what "average" means. All the arguments about weak links, variance, difficulty recruiting depth that allows DL to rotate, etc. apply to the majority of the Pac-12 as well. There is no way you can slice it - other than actual performance in college - that supports the idea Stanford has had below-average DL talent at any point under Shaw. Half the conference would trade for our DL room at the time of recruitment and NFL evaluators would pick Stanford's DL over half the conference at the NFL combine stage. That's really the bottom line.

While brevity is not my thing, if I had to summarize my view of our DL situation it would go like this: without talent like Solomon Thomas/Aziz Shittu and without coaching like Randy Hart, we fell off a cliff; the recruiting drop-off should not be ignored, but it's about not getting the difference makers (i.e., DL1) rather than the volume/variance arguments you're making (i.e., DL3/DL4). If you look at other Pac-12 teams the "average" DL3/DL4 situation is no prettier than ours; regardless of how much stock one puts in the recruiting, the evidence is overwhelming that our defense massively under-performs its talent, whether you look at recruiting or NFL evaluations.

[msqueri]

DL has had insufficient numbers from the time Diron Took over until last year’s class.  He recruited only 2 in 2017 and 2018, 1 in 2019, and 2 again in 2020. Things could have worked out better if the DLs that did come panned out at an excellent rate. Instead, Pakola basically never played and Fox and Keneley proved too skinny to fill a 3-4 DE spot. To boot, RJ and DWP weren’t great players and never matched their recruiting profile. Plus the 2020 guys were lightly recruited. The result is ONE really good DL in a 4 year window.

2017 had two (Johnson and Wade-Perry), 2018 had two (Andres Fox and Thomas Booker), 2019 had 1 (Pakola), and 2020 had three (Phillips, Keneley, Franklin). Other teams also have washouts (especially at DL) and also have front seven big athletes/tweeners who move around or end up not having the body type for a particular position. Even then, that 2020 trio now includes two guys slotted for interior DL and another for a position we're signaling will sometimes look like OLB and sometimes like like DE.

If we're going to call eight DL in a four year stretch insufficient, I'd like to compare it to past periods when we did recruit DL sufficiently. Perhaps the four Harbaugh classes (2008-2011) would be an interesting comparison. That was Padric Scott, Terrence Stephens, Josh Mauro, Ben Gardner, Henry Anderson, Eddie Plantaric, Anthony Hayes, Lance Callihan, JB Salem, and Charlie Hopkins. That is two more DL than we recruited in 2017-2020, and the Harbaugh group had many more washouts/useless guys. A good case can be made that we added such a h.u.g.e 2011 DL class (Hopkins, Hayes, Salem, and Callihan) because Scott and Plantaric were duds, that Salem and Callihan were the type of Plan B/C reaches Shaw has avoided, and that ultimately the four man 2011 DL class did absolutely nothing for Stanford football. If anything, it's an example of how volume for the sake of volume gets you little.

Instead, a picture emerges in which the difference between the way that previous recruiting panned out compared to the way the more recent recruiting panned out was that Hart and Turley (and the scheme of Fangio and Mason) were able to get much more out of Stephens, Mauro, Gardner, and Anderson than Reynolds and Carroll (and Anderson's scheme) were able to get out of Johnson, Wade-Perry, Fox, and Booker. That wasn't a recruiting problem. We were over-the-moon thrilled with Johnson, Wade-Perry, Fox, and Booker in back-to-back classes. One can argue that it was partly an evaluation issue, but Booker playing so much and so well early is a data point supportive of my argument that DL simply have not developed under Reynolds. The Harbaugh-era DL, essentially to a man (Mauro had his big emergence in his fourth year), all had big-time breakthroughs in their third season. To put it mildly that has not been the case with Reynolds' charges (I have a separate analysis showing the lack of late career improvement by DL in recent years).

I can't believe Shaw/Anderson/Reynolds are turning me into more of an Xs and Os guy. That's how bad their coaching is.

[msqueri]

DL has had insufficient numbers from the time Diron Took over until last year’s class.  He recruited only 2 in 2017 and 2018, 1 in 2019, and 2 again in 2020. Things could have worked out better if the DLs that did come panned out at an excellent rate. Instead, Pakola basically never played and Fox and Keneley proved too skinny to fill a 3-4 DE spot. To boot, RJ and DWP weren’t great players and never matched their recruiting profile. Plus the 2020 guys were lightly recruited. The result is ONE really good DL in a 4 year window.

2017 had two (Johnson and Wade-Perry), 2018 had two (Andres Fox and Thomas Booker), 2019 had 1 (Pakola), and 2020 had three (Phillips, Keneley, Franklin). Other teams also have washouts (especially at DL) and also have front seven big athletes/tweeners who move around or end up not having the body type for a particular position. Even then, that 2020 trio now includes two guys slotted for interior DL and another for a position we're signaling will sometimes look like OLB and sometimes like like DE.

If we're going to call eight DL in a four year stretch insufficient, I'd like to compare it to past periods when we did recruit DL sufficiently. Perhaps the four Harbaugh classes (2008-2011) would be an interesting comparison. That was Padric Scott, Terrence Stephens, Josh Mauro, Ben Gardner, Henry Anderson, Eddie Plantaric, Anthony Hayes, Lance Callihan, JB Salem, and Charlie Hopkins. That is two more DL than we recruited in 2017-2020, and the Harbaugh group had many more washouts/useless guys. A good case can be made that we added such a h.u.g.e 2011 DL class (Hopkins, Hayes, Salem, and Callihan) because Scott and Plantaric were duds, that Salem and Callihan were the type of Plan B/C reaches Shaw has avoided, and that ultimately the four man 2011 DL class did absolutely nothing for Stanford football. If anything, it's an example of how volume for the sake of volume gets you little.

Instead, a picture emerges in which the difference between the way that previous recruiting panned out compared to the way the more recent recruiting panned out was that Hart and Turley (and the scheme of Fangio and Mason) were able to get much more out of Stephens, Mauro, Gardner, and Anderson than Reynolds and Carroll (and Anderson's scheme) were able to get out of Johnson, Wade-Perry, Fox, and Booker. That wasn't a recruiting problem. We were over-the-moon thrilled with Johnson, Wade-Perry, Fox, and Booker in back-to-back classes. One can argue that it was partly an evaluation issue, but Booker playing so much and so well early is a data point supportive of my argument that DL simply have not developed under Reynolds. The Harbaugh-era DL, essentially to a man (Mauro had his big emergence in his fourth year), all had big-time breakthroughs in their third season. To put it mildly that has not been the case with Reynolds' charges (I have a separate analysis showing the lack of late career improvement by DL in recent years).

I can't believe Shaw/Anderson/Reynolds are turning me into more of an Xs and Os guy. That's how bad their coaching is.

[TrumpCard]

Although I agree with almost everything above, I want to note that Reynolds, the DL, and the Stanford defense in general might have been better in 2017 than we're acknowledging.  The defense did give up too many points in early games against UCLA and USC, but note that those team both had top 10 NFL draft picks at QB that year.  Also, Jackson and Cotton were rounding into form by midseason, and I was cautiously optimistic.  Even after Cotton went down in the close loss to WSU, the DL and the defense were pretty resilient and won two games against top 10 teams (UW and ND).  Stanford ended the regular season 9-3 and gave up more than 24 points only twice in those two early games against elite QBs.  If not for Cotton's injury, I think Stanford's defense would not have fallen apart in the post-season, and I think the team would have won the Pac-12 title, or if not, at least the team would have won the Alamo Bowl.  Going 1-1 in the post-season would have meant a 10-win season and a top-15 ranking in the final polls.  That would have changed the narrative a lot, and I think the defensive stats would have looked better too.

[TrumpCard]

Although I agree with almost everything above, I want to note that Reynolds, the DL, and the Stanford defense in general might have been better in 2017 than we're acknowledging.  The defense did give up too many points in early games against UCLA and USC, but note that those team both had top 10 NFL draft picks at QB that year.  Also, Jackson and Cotton were rounding into form by midseason, and I was cautiously optimistic.  Even after Cotton went down in the close loss to WSU, the DL and the defense were pretty resilient and won two games against top 10 teams (UW and ND).  Stanford ended the regular season 9-3 and gave up more than 24 points only twice in those two early games against elite QBs.  If not for Cotton's injury, I think Stanford's defense would not have fallen apart in the post-season, and I think the team would have won the Pac-12 title, or if not, at least the team would have won the Alamo Bowl.  Going 1-1 in the post-season would have meant a 10-win season and a top-15 ranking in the final polls.  That would have changed the narrative a lot, and I think the defensive stats would have looked better too.

[msqueri]

Although I agree with almost everything above, I want to note that Reynolds, the DL, and the Stanford defense in general might have been better in 2017 than we're acknowledging.  The defense did give up too many points in early games against UCLA and USC, but note that those team both had top 10 NFL draft picks at QB that year.  Also, Jackson and Cotton were rounding into form by midseason, and I was cautiously optimistic.  Even after Cotton went down in the close loss to WSU, the DL and the defense were pretty resilient and won two games against top 10 teams (UW and ND).  Stanford ended the regular season 9-3 and gave up more than 24 points only twice in those two early games against elite QBs.  If not for Cotton's injury, I think Stanford's defense would not have fallen apart in the post-season, and I think the team would have won the Pac-12 title, or if not, at least the team would have won the Alamo Bowl.  Going 1-1 in the post-season would have meant a 10-win season and a top-15 ranking in the final polls.  That would have changed the narrative a lot, and I think the defensive stats would have looked better too.

Good observations. The best way I know of to assess single-game effectiveness and trends over time, at least in a way that accounts for strength of competition and such, is to look at yards per play allowed compared to what the opposing offense typically mustered. Here's what 2017 looked like:

Rice - 3.89 (1.55 better than Rice's average)
USC - 8.42 (1.87 worse than USC's average)
San Diego State - 4.53 (1.76 better than San Diego State's average)
UCLA - 7.93 (1.54 worse than UCLA's average)
Arizona State - 5.76 (0.05 worse than Arizona State's average)
Utah - 5.23 (0.41 better than Utah's average)
Oregon - 5.52 (0.58 better than Oregon's average)
Oregon State - 4.19 (0.99 better than Oregon State's average)
Washington State - 5.58 (0.02 worse than Washington State's average)
Washington - 5.80 (0.62 better than Washington's average)
Cal - 5.91 (0.74 worse than Cal's average)
Notre Dame - 5.68 (0.71 better than Notre Dame's average)
USC - 7.48 (0.93 worse than USC's average)
TCU - 7.28 (1.19 worse than TCU's average)

Exactly half the games were better than defenses tended to do against the opponent and half worse. Cotton got hurt early in the Notre Dame game and at the time I certainly shared your sense that that hurt the defense in the Pac-12 championship and Alamo Bowl, both of which were not good defensive performances compared to those opposing offenses' norms. Then again, USC roasted us even worse early in the season compared to the Pac-12 championship game. All in all, I think the verdict of the advanced metrics I cited (#52 in Defensive FEI) is a fair representation of the quality of that defense. It was a bit above-average, I think. Considering there were 7-9 NFL players (depending on how you count guys who briefly made rosters but never played in the league) who all played 200+ snaps that year.....yikes, holy underachievement, Batman.

On the other point of how close the narrative was to being much different.....yes, great point. There were a lot of things that year that if they had gone differently we're talking about an indisputably great season. Make Costello QB1 at the beginning of the season and it's hard to see San Diego State being a loss. Cotton's injury, as you mentioned. The goal line heartbreak against USC in the Pac-12 championship. Justin Reid getting hurt in the Alamo Bowl. And many more I'm not remembering. Easily could have been a 10+ win season. That being said, for what it's worth I do view that as a great season and it's not part of my narrative of Shaw's overall decline except in the sense that it's evidence of defensive decline that such a talented defense was only a bit above-average. That team was really good overall, as was the TCU team we played in the bowl, and it easily could have been a 10 win season. 2018 is where to demarcate the real decline. But these observations on defensive decline we're all making in this thread speak to the fact there were warning seasons before the chickens came home to roost. 2015 was one of our greatest all-time seasons and 2017 was pretty high up there too yet both presented evidence of troubling defensive decline.

[msqueri]

Although I agree with almost everything above, I want to note that Reynolds, the DL, and the Stanford defense in general might have been better in 2017 than we're acknowledging.  The defense did give up too many points in early games against UCLA and USC, but note that those team both had top 10 NFL draft picks at QB that year.  Also, Jackson and Cotton were rounding into form by midseason, and I was cautiously optimistic.  Even after Cotton went down in the close loss to WSU, the DL and the defense were pretty resilient and won two games against top 10 teams (UW and ND).  Stanford ended the regular season 9-3 and gave up more than 24 points only twice in those two early games against elite QBs.  If not for Cotton's injury, I think Stanford's defense would not have fallen apart in the post-season, and I think the team would have won the Pac-12 title, or if not, at least the team would have won the Alamo Bowl.  Going 1-1 in the post-season would have meant a 10-win season and a top-15 ranking in the final polls.  That would have changed the narrative a lot, and I think the defensive stats would have looked better too.

Good observations. The best way I know of to assess single-game effectiveness and trends over time, at least in a way that accounts for strength of competition and such, is to look at yards per play allowed compared to what the opposing offense typically mustered. Here's what 2017 looked like:

Rice - 3.89 (1.55 better than Rice's average)
USC - 8.42 (1.87 worse than USC's average)
San Diego State - 4.53 (1.76 better than San Diego State's average)
UCLA - 7.93 (1.54 worse than UCLA's average)
Arizona State - 5.76 (0.05 worse than Arizona State's average)
Utah - 5.23 (0.41 better than Utah's average)
Oregon - 5.52 (0.58 better than Oregon's average)
Oregon State - 4.19 (0.99 better than Oregon State's average)
Washington State - 5.58 (0.02 worse than Washington State's average)
Washington - 5.80 (0.62 better than Washington's average)
Cal - 5.91 (0.74 worse than Cal's average)
Notre Dame - 5.68 (0.71 better than Notre Dame's average)
USC - 7.48 (0.93 worse than USC's average)
TCU - 7.28 (1.19 worse than TCU's average)

Exactly half the games were better than defenses tended to do against the opponent and half worse. Cotton got hurt early in the Notre Dame game and at the time I certainly shared your sense that that hurt the defense in the Pac-12 championship and Alamo Bowl, both of which were not good defensive performances compared to those opposing offenses' norms. Then again, USC roasted us even worse early in the season compared to the Pac-12 championship game. All in all, I think the verdict of the advanced metrics I cited (#52 in Defensive FEI) is a fair representation of the quality of that defense. It was a bit above-average, I think. Considering there were 7-9 NFL players (depending on how you count guys who briefly made rosters but never played in the league) who all played 200+ snaps that year.....yikes, holy underachievement, Batman.

On the other point of how close the narrative was to being much different.....yes, great point. There were a lot of things that year that if they had gone differently we're talking about an indisputably great season. Make Costello QB1 at the beginning of the season and it's hard to see San Diego State being a loss. Cotton's injury, as you mentioned. The goal line heartbreak against USC in the Pac-12 championship. Justin Reid getting hurt in the Alamo Bowl. And many more I'm not remembering. Easily could have been a 10+ win season. That being said, for what it's worth I do view that as a great season and it's not part of my narrative of Shaw's overall decline except in the sense that it's evidence of defensive decline that such a talented defense was only a bit above-average. That team was really good overall, as was the TCU team we played in the bowl, and it easily could have been a 10 win season. 2018 is where to demarcate the real decline. But these observations on defensive decline we're all making in this thread speak to the fact there were warning seasons before the chickens came home to roost. 2015 was one of our greatest all-time seasons and 2017 was pretty high up there too yet both presented evidence of troubling defensive decline.

[TrumpCard]

Question for those who follow these things more closely than I do nowadays:
What is the current consensus on the pecking order of the services that evaluate college football recruits?
I recall ESPN used to be questionable, and place too much emphasis on speed over other skills/traits.  Is that still the case?
Is On3 any good?  
Does anyone pay attention to Rank by Offers anymore or have offers proliferated and become so hard to decipher that the ranking system lost value?  
Has the bias in favor of recruits who attend camps increased, decreased, or remained constant at Rivals and 247?
Do ESPN and On3 have affiliations with camp series?
Do you still like looking at the composite to get a balanced view of the recruit's profile, or are some of the services so skewed that you'd rather just look at one or two of the rankings instead of the composite?
Thanks in advance for your input.

[TrumpCard]

Question for those who follow these things more closely than I do nowadays:
What is the current consensus on the pecking order of the services that evaluate college football recruits?
I recall ESPN used to be questionable, and place too much emphasis on speed over other skills/traits.  Is that still the case?
Is On3 any good?  
Does anyone pay attention to Rank by Offers anymore or have offers proliferated and become so hard to decipher that the ranking system lost value?  
Has the bias in favor of recruits who attend camps increased, decreased, or remained constant at Rivals and 247?
Do ESPN and On3 have affiliations with camp series?
Do you still like looking at the composite to get a balanced view of the recruit's profile, or are some of the services so skewed that you'd rather just look at one or two of the rankings instead of the composite?
Thanks in advance for your input.

[BobK]

247 is the best.

[BobK]

247 is the best.

[SamuelMcF]

247 is the best, particularly with their own evaluations.

I like that On3's consensus weights all 4 services equally, and their industry comparison shows all 4 services' rankings side by side.
(E.g. here's Sedrick Irvin, who is a 4* by the 3 other services that aren't On3: https://www.on3.com/db/sedrick-irvin-jr....omparison/)

I have no idea how On3's prediction algorithm or NIL valuation work, so I put no stock into them whatsoever. No idea on their internal evaluations either as they are brand new.

Rivals is ok but weighs camp participation too highly, and sometimes gets positions wrong. (They still think Elijah Hughes is an offensive guard, for example.) Rating recruits on a weird 5.3-6.1 scale makes no sense and isn't nearly granular enough.

ESPN is all over the place and heavily east coast biased, and I generally don't take their rankings seriously.

I have never paid attention to Rank By Offers, and completely forgot it existed until you mentioned it. (I will still not pay attention to it.)

[SamuelMcF]

247 is the best, particularly with their own evaluations.

I like that On3's consensus weights all 4 services equally, and their industry comparison shows all 4 services' rankings side by side.
(E.g. here's Sedrick Irvin, who is a 4* by the 3 other services that aren't On3: https://www.on3.com/db/sedrick-irvin-jr....omparison/)

I have no idea how On3's prediction algorithm or NIL valuation work, so I put no stock into them whatsoever. No idea on their internal evaluations either as they are brand new.

Rivals is ok but weighs camp participation too highly, and sometimes gets positions wrong. (They still think Elijah Hughes is an offensive guard, for example.) Rating recruits on a weird 5.3-6.1 scale makes no sense and isn't nearly granular enough.

ESPN is all over the place and heavily east coast biased, and I generally don't take their rankings seriously.

I have never paid attention to Rank By Offers, and completely forgot it existed until you mentioned it. (I will still not pay attention to it.)

[TrumpCard]

Thanks, Samuel and Bob.  It's probably good for Stanford that 247's internal evaluations are seen as the most accurate.  247 has Irvin, Pale, and Bachmeier all as 4-star recruits, with only Borden as a 3-star where the consensus says 4-star.  Among the uncommitted recruits, it looks like 247 has Minch, Snowden, and Brandt as 4-stars, and only Link as a 3-star where the consensus says 4-star.  So my guess is that 247 will have a more positive view of Stanford's final class than the consensus/composite.  (I'm ignoring the differences within the star categories as those tend to be less material to class rankings, but even in those instances, it seems like 247 may be slightly more favorable to Stanford recruits on balance.)

[TrumpCard]

Thanks, Samuel and Bob.  It's probably good for Stanford that 247's internal evaluations are seen as the most accurate.  247 has Irvin, Pale, and Bachmeier all as 4-star recruits, with only Borden as a 3-star where the consensus says 4-star.  Among the uncommitted recruits, it looks like 247 has Minch, Snowden, and Brandt as 4-stars, and only Link as a 3-star where the consensus says 4-star.  So my guess is that 247 will have a more positive view of Stanford's final class than the consensus/composite.  (I'm ignoring the differences within the star categories as those tend to be less material to class rankings, but even in those instances, it seems like 247 may be slightly more favorable to Stanford recruits on balance.)

[msqueri]

There is a cyclical aspect to this as well. Rivals used to be the best, then it was 247, and it remains to be seen whether/how dominant On3 becomes. There may be a simple explanation for this: all three were essentially created by the same guy. He goes all in on creating a network of college sports fan sites and then sells them to a major media company. Rinse and repeat. He's been doing it for decades. He sold his first recruiting service, Alliance Sports, to Rival Networks for $3 million in 2000. When Rival Networks went bankrupt, he rebought the assets for $500,000. Then he sold Rivals to Yahoo in 2007 for $98 million. As soon as his non-compete ended he created 247 and then five years later he sold 247 to CBS for what's believed to be upward of $200 million. And now he's launched On3. The smart money is probably on On3 becoming the dominant player within the next five years.

247 gets much more traffic and, presumably, more revenue than Rivals, which leads to more resources. Theoretically that should lead to higher quality and more predictive rankings, though I'm not aware of all that many good comparisons across the services, and any comparisons there are would provide retrospective/lagging insight. For instance, in 2017 people crunched the NFL draft results to find Rivals was the best at predicting first round NFL draft picks, but that tells you something about recruiting rankings circa 2012-2013. A lot changed since then, namely 247 eating Rivals' lunch.

ESPN is completely worthless. If you pay attention to their rankings you'll notice they are a) heavily imbalanced toward players from the Southeast (and I am one who thinks the Southeast is BY FAR better at football than other regions, but there can be no excuse for how many of ESPN's top recruits tend to be from there) and b) heavily biased in favor of players that happen to agree to play in the Under Armour All-America Game, which is televised by ESPN. ESPN runs its recruiting work (can't call it a "service" with a straight face) on the cheap, knowing that its brand dominance generally in sports will get it eyeballs no matter how useless or incredible its insights actually are. And, of course, recruiting is way down the hierarchy of things core to its business, very much unlike 247/Rivals/On3.

I love the idea of a composite, both 247 and On3. For me that's much more meaningful than a single service's (even 247) evaluation. I don't yet have an opinion on whether 247 or On3's composite is better. I will probably trust On3's composite, Recruiting Prediction Machine, and NIL valuations much more several years from now than I do now, as they will presumably learn/adapt as an organization, gain more resources, and plug past data into their machine learning.

I don't share SamuelMcF's dislike of the Rivals scale. While I don't understand why the numbers they pick are in the range they are/capped at 6.1, I think the instinct is very sound to find a way to distinguish between recruits that isn't so overly precise as an exact numerical ranking (what kind of confidence can one have that the #611 prospect is better than the #683 prospect? None) or as blunt and un-nuanced as a star rating. Fans who put a lot of stock in low 4-star rankings and/or ignore the signals sent by high 3-star rankings have some of the most misguided views in recruiting. To me, the Rivals' approach finds a defensible middle ground. As you can see in my three-dimensional analyses, I value the opportunity to pick out comparable recruits in the view of a major recruiting service. That would be a much less valuable exercise if I did so based on 3/4/5 stars and if I tried to come up with my own range of which increments are most comparable for given recruits that would be subjective and require work Rivals' approach makes unnecessary.

I thought Rank By Offers was flawed but I appreciated some attempt to algorithmically compare offer quality. I notice all the time Stanford fans citing an offer from _____ (insert decent football program here, or insert hard-nosed identity program with respect to an OL recruit, or so forth) as if that's supposed to tell us something meaningful about a recruit. That totally ignores how many offers are provisional/not followed through on as committable offers and how much offer proliferation there has been generally. It was nice to have a ranking to try to generally, if imperfectly, compare offer strength. But Rank By Offers has been defunct for five years.

On whether 247 having credibility helps our image, bear in mind that the main rankings 247 uses for individuals and team rankings are the composite, not the 247 rankings. I'm not sure I've noticed one service systematically liking Stanford recruits more/less than the others over multiple years but within individual years there definitely can be a divergence in views on how our recruits are perceived.

[msqueri]

There is a cyclical aspect to this as well. Rivals used to be the best, then it was 247, and it remains to be seen whether/how dominant On3 becomes. There may be a simple explanation for this: all three were essentially created by the same guy. He goes all in on creating a network of college sports fan sites and then sells them to a major media company. Rinse and repeat. He's been doing it for decades. He sold his first recruiting service, Alliance Sports, to Rival Networks for $3 million in 2000. When Rival Networks went bankrupt, he rebought the assets for $500,000. Then he sold Rivals to Yahoo in 2007 for $98 million. As soon as his non-compete ended he created 247 and then five years later he sold 247 to CBS for what's believed to be upward of $200 million. And now he's launched On3. The smart money is probably on On3 becoming the dominant player within the next five years.

247 gets much more traffic and, presumably, more revenue than Rivals, which leads to more resources. Theoretically that should lead to higher quality and more predictive rankings, though I'm not aware of all that many good comparisons across the services, and any comparisons there are would provide retrospective/lagging insight. For instance, in 2017 people crunched the NFL draft results to find Rivals was the best at predicting first round NFL draft picks, but that tells you something about recruiting rankings circa 2012-2013. A lot changed since then, namely 247 eating Rivals' lunch.

ESPN is completely worthless. If you pay attention to their rankings you'll notice they are a) heavily imbalanced toward players from the Southeast (and I am one who thinks the Southeast is BY FAR better at football than other regions, but there can be no excuse for how many of ESPN's top recruits tend to be from there) and b) heavily biased in favor of players that happen to agree to play in the Under Armour All-America Game, which is televised by ESPN. ESPN runs its recruiting work (can't call it a "service" with a straight face) on the cheap, knowing that its brand dominance generally in sports will get it eyeballs no matter how useless or incredible its insights actually are. And, of course, recruiting is way down the hierarchy of things core to its business, very much unlike 247/Rivals/On3.

I love the idea of a composite, both 247 and On3. For me that's much more meaningful than a single service's (even 247) evaluation. I don't yet have an opinion on whether 247 or On3's composite is better. I will probably trust On3's composite, Recruiting Prediction Machine, and NIL valuations much more several years from now than I do now, as they will presumably learn/adapt as an organization, gain more resources, and plug past data into their machine learning.

I don't share SamuelMcF's dislike of the Rivals scale. While I don't understand why the numbers they pick are in the range they are/capped at 6.1, I think the instinct is very sound to find a way to distinguish between recruits that isn't so overly precise as an exact numerical ranking (what kind of confidence can one have that the #611 prospect is better than the #683 prospect? None) or as blunt and un-nuanced as a star rating. Fans who put a lot of stock in low 4-star rankings and/or ignore the signals sent by high 3-star rankings have some of the most misguided views in recruiting. To me, the Rivals' approach finds a defensible middle ground. As you can see in my three-dimensional analyses, I value the opportunity to pick out comparable recruits in the view of a major recruiting service. That would be a much less valuable exercise if I did so based on 3/4/5 stars and if I tried to come up with my own range of which increments are most comparable for given recruits that would be subjective and require work Rivals' approach makes unnecessary.

I thought Rank By Offers was flawed but I appreciated some attempt to algorithmically compare offer quality. I notice all the time Stanford fans citing an offer from _____ (insert decent football program here, or insert hard-nosed identity program with respect to an OL recruit, or so forth) as if that's supposed to tell us something meaningful about a recruit. That totally ignores how many offers are provisional/not followed through on as committable offers and how much offer proliferation there has been generally. It was nice to have a ranking to try to generally, if imperfectly, compare offer strength. But Rank By Offers has been defunct for five years.

On whether 247 having credibility helps our image, bear in mind that the main rankings 247 uses for individuals and team rankings are the composite, not the 247 rankings. I'm not sure I've noticed one service systematically liking Stanford recruits more/less than the others over multiple years but within individual years there definitely can be a divergence in views on how our recruits are perceived.

[old spanish trail]

Square, you have more information than the rest of us put together. Thanks.

[old spanish trail]

Square, you have more information than the rest of us put together. Thanks.

[Phogge]

Hey Sam, any idea how the staff is handling recruiting in the wake of the departure of UCLA/SC? If I'm a kid or a parent I would be asking a bunch of questions re the (unknown) future. If they don't have a plausible answer I'd start looking elsewhere.

[Phogge]

Hey Sam, any idea how the staff is handling recruiting in the wake of the departure of UCLA/SC? If I'm a kid or a parent I would be asking a bunch of questions re the (unknown) future. If they don't have a plausible answer I'd start looking elsewhere.

[TrumpCard]

Thanks for the input everyone.  Very helpful.  

FWIW, if you just want to compare the 247 class rankings to the Rivals class rankings, then 247 is almost always more generous to Stanford.
I recognize this may be the result of the method for ranking classes rather than the evaluation of individual players.  The difference in class rankings may just occur, for example, because Rivals caps out after the top 20 players and implicitly penalizes small classes.  I don't have detailed player-by-player information to say that 247 evaluates Stanford recruits as being better than Rivals does.  Here's the class ranking for each year with Rivals first and 247 second:

2022: 25, 19
2021: 46, 43
2020: 23, 21
2019: 24, 19
2018: 57, 40
2017: 19, 14
2016: 19, 16
2015: 18, 24 **This is the one exception in the last 10 years
2014: 14, 13
2013: 62, 52

[TrumpCard]

Thanks for the input everyone.  Very helpful.  

FWIW, if you just want to compare the 247 class rankings to the Rivals class rankings, then 247 is almost always more generous to Stanford.
I recognize this may be the result of the method for ranking classes rather than the evaluation of individual players.  The difference in class rankings may just occur, for example, because Rivals caps out after the top 20 players and implicitly penalizes small classes.  I don't have detailed player-by-player information to say that 247 evaluates Stanford recruits as being better than Rivals does.  Here's the class ranking for each year with Rivals first and 247 second:

2022: 25, 19
2021: 46, 43
2020: 23, 21
2019: 24, 19
2018: 57, 40
2017: 19, 14
2016: 19, 16
2015: 18, 24 **This is the one exception in the last 10 years
2014: 14, 13
2013: 62, 52

[TrumpCard]

Just to temper everyone's expectations, three CBs for Lonergan to Bama tonight...including Steve Wiltfong.  Not good.

[TrumpCard]

Just to temper everyone's expectations, three CBs for Lonergan to Bama tonight...including Steve Wiltfong.  Not good.

[Phogge]

Why would the kid come to Stanford with the PAC about to fall apart?

[Phogge]

Why would the kid come to Stanford with the PAC about to fall apart?

[BobK]

Or it means Lonergan won’t be picking South Carolina ?? Could be.

[BobK]

Or it means Lonergan won’t be picking South Carolina ?? Could be.

[CardinalSagehen]

Here are a couple stories from the last 24 hours on Lonergan's plans:

https://www.si.com/college/stanford/recr...to-alabama

https://247sports.com/college/stanford/A...189831050/

If Lonergan does in fact commit to Alabama, it'll mean that every one of the five quarterbacks Stanford has offered in 2023 will have committed elsewhere.

Getting "Fong'd" [Steve Wiltfong prediction] usually means the projected destination is correct.

It ain't over 'till it's over, but this isn't looking good for the good guys. Stanford's QB room is going to be totally uncrowded - I hope he considers that.  But with the opportunity to transfer at any time, he probably just figures that he'll have the option to transfer if he doesn't win the starting job at Alabama.

[CardinalSagehen]

Here are a couple stories from the last 24 hours on Lonergan's plans:

https://www.si.com/college/stanford/recr...to-alabama

https://247sports.com/college/stanford/A...189831050/

If Lonergan does in fact commit to Alabama, it'll mean that every one of the five quarterbacks Stanford has offered in 2023 will have committed elsewhere.

Getting "Fong'd" [Steve Wiltfong prediction] usually means the projected destination is correct.

It ain't over 'till it's over, but this isn't looking good for the good guys. Stanford's QB room is going to be totally uncrowded - I hope he considers that.  But with the opportunity to transfer at any time, he probably just figures that he'll have the option to transfer if he doesn't win the starting job at Alabama.

[MVDan]

I wouldn’t give up on a Stanford QB yet if Lonergan doesn’t come next year. Besides Patu, Ashton Daniels is here as well. His trajectory reminds me a lot of Kevin Hogan and is a mobile strong armed QB that just seems to win when he plays.

[MVDan]

I wouldn’t give up on a Stanford QB yet if Lonergan doesn’t come next year. Besides Patu, Ashton Daniels is here as well. His trajectory reminds me a lot of Kevin Hogan and is a mobile strong armed QB that just seems to win when he plays.