Football Outsiders Quick Reads on Josh Allen

[TheFunPolice]

2 hours ago, klos63 said:

typical buffalo - please like us.  I don't know what you watch, but I've seen endless praise for Allen. Many picked the Bills to win the division.  But that doesn't fit the ' I don't get no respect' narrative.

 

Patriots used the same thing every year.  People would predict them to go 14-2 and they would be offended that they were picked to lose 2 games. LOL

 

Football players LOVE this stuff. Hey, whatever works!

[Clyde Smith]

2 hours ago, klos63 said:

typical buffalo - please like us.  I don't know what you watch, but I've seen endless praise for Allen. Many picked the Bills to win the division.  But that doesn't fit the ' I don't get no respect' narrative.

Relax dude, the team is winning. That phrase was from Jerry Hughes, not me but I like it. Did you watch the video clip before quoting me?

[x-BillzeBubba]

here's some other statistical analysis from the way-back machine...lol

 

"If this works, then stats really are for losers, I guess."

 

https://www.sbnation.com/nfl/2018/4/24/17271686/josh-allen-nfl-draft-2018-stats-analysis-comparisons

Edited September 22, 2020 by x-BillzeBubba

[Bleeding Bills Blue]

31 minutes ago, JohnNord said:

I love this paragraph.  Aaron Schatz and FO were dunked on so bad their last past analysis of Allen.  This is their way of acknowledging that they were probably wrong, without ever saying it.

 

“We'll start with Josh Allen and the Bills. The consensus at Football Outsiders has always been pessimistic about Allen's chances for success. Our QBASE article in 2018 noted Allen's "horrifying" statistics at the University of Wyoming, and we referred to him as "a parody of an NFL quarterback prospect" in Football Outsiders Almanac that same year.”

 

It's tough.  The comparisons in the last 20 years of drafted prospects are guys like Bortles, Kaepernick, Locker and gabbert.  Maybe to a lesser extent Cutler or Stafford - big arms and lower completion percentages in college.  His numbers from an accuracy perspective were worse than all of them - so naturally from a statistical mindset you're looking at a QB who has to massively improve just to hit the minimum baseline, and there aren't really a ton of examples of that happening.  

 

I think examples you can look at if you go back further (look at it a bit more like 5% improvement in 1995 is more significant than 5% today), might be Mcnair or Favre from an arm talent perspective.  They were inaccurate in college (sub 60%) and showed improvement at the NFL level.  I think you would've liked to see something more like Mcnairs 4th year (like 6000 yards and 50+ TDs) at Wyoming - but the talent surrounding him was not very good, and consider Mcnair was playing in D1-AA.  Looking at Favre's college numbers, nothing there indicates a player who would at some point complete 68% of his passes in a year, or lead the NFL in TDs. 

I don't think FO is trying to find the next Favre when they look at these types of analysis.  The trend indicates you are getting a player like the 4 above - and that trend would indicate failure.  But in all honestly - Allen reminds me a bit of Favre/Mcnair with how he plays.  There's the reckless confidence of a favre, buying time and scrambling like mcnair, rocket arms.  

[SoTier]

 

30 minutes ago, JohnNord said:

I love this paragraph.  Aaron Schatz and FO were dunked on so bad their last past analysis of Allen.  This is their way of acknowledging that they were probably wrong, without ever saying it.

 

“We'll start with Josh Allen and the Bills. The consensus at Football Outsiders has always been pessimistic about Allen's chances for success. Our QBASE article in 2018 noted Allen's "horrifying" statistics at the University of Wyoming, and we referred to him as "a parody of an NFL quarterback prospect" in Football Outsiders Almanac that same year.”

 

I disagree that FO isn't acknowledging that they were probably wrong about Allen.   They put their skepticism about Allen coming out of college at the very front and center of their article, and they concluded that even if he doesn't do as well going forward, that he's at the very least developed into a decent NFL starter capable of taking the Bills to some success in the playoffs.   

 

The reality is that Allen was not a very good prospect coming out of college.  He had a rocket arm, size, and running ability but he lacked almost all the tangible skills QBs need to succeed in the NFL -- and his rookie season really only provided a glimmer of hope that he could improve.   My guess is that Allen suffered from a serious lack of competent coaching/mentoring on both the collegiate and pro levels until Ken Dorsey took him in hand last season.   Allen got better as the season went on last year, which was an encouraging sign, and he appears to have significantly  improved over the off season despite the disruptions of the normal preseason routines because of the coronavirus pandemic.  

 

A big test comes this Sunday.  The Rams aren't the Jets or the Carp.

[billsfan89]

4 hours ago, HappyDays said:

What PFF does isn't analytics. They are subjectively grading players. Football Outsiders just takes raw data and contextualizes it for game situation and opponent. In their system a 5 yard pass on 3rd and 4 is graded better than a 10 yard pass on 3rd and 20, even though the 10 yard pass looks better in traditional statistics.

 

PFF does try to do situation weighing but for whatever reason they don't do a good enough job of it. Football Outsiders grades every player on every play like PFF but their metrics and other subjective elements are just better than PFF. I don't know if PFF has lesser qualified people watching the games or what their deal is but their ratings always seem more off.

[billsfan89]

5 minutes ago, SoTier said:

 

 

I disagree that FO isn't acknowledging that they were probably wrong about Allen.   They put their skepticism about Allen coming out of college at the very front and center of their article, and they concluded that even if he doesn't do as well going forward, that he's at the very least developed into a decent NFL starter capable of taking the Bills to some success in the playoffs.   

 

The reality is that Allen was not a very good prospect coming out of college.  He had a rocket arm, size, and running ability but he lacked almost all the tangible skills QBs need to succeed in the NFL -- and his rookie season really only provided a glimmer of hope that he could improve.   My guess is that Allen suffered from a serious lack of competent coaching/mentoring on both the collegiate and pro levels until Ken Dorsey took him in hand last season.   Allen got better as the season went on last year, which was an encouraging sign, and he appears to have significantly  improved over the off season despite the disruptions of the normal preseason routines because of the coronavirus pandemic.  

 

A big test comes this Sunday.  The Rams aren't the Jets or the Carp.

 

There was very legitimate concerns to be had about Josh coming out of college. Accuracy is something that can rarely be taught at the pro level and Josh also had to develop his footwork and decision making game awareness. Josh also came from a small college and rarely played against great competition. The Bills gambled on Josh having elite throwing power, running ability, size and strength and his intangibles as a smart leader making him able to harness those abilities at a pro-level. 

 

The Bills also gambled that his accuracy issues while there were overstated in college due to the deep passing scheme and low talent level would deflate completion percentage. The Bills so far appear to have been correct, experience and coaching have helped his footwork, decision making and game processing/awareness. Which has tightened up his accuracy and allowed him to harness his elite attributes thus far in 2020. 

 

So I don't blame PFF or FO for not liking Josh out of college. There have been a lot of other "million dollar arm" type QB's who never developed accuracy in the pros. Josh had a good amount of red flags. But after 2019 Josh gave hints of being able to win and improve, you had to at least upgrade your outlook of him somewhat. And thusfar in 2020 he has only shown to progress more and is likely at the very least a mid-level starting caliber QB.

[eball]

23 minutes ago, billsfan89 said:

 

There was very legitimate concerns to be had about Josh coming out of college. Accuracy is something that can rarely be taught at the pro level and Josh also had to develop his footwork and decision making game awareness. Josh also came from a small college and rarely played against great competition. The Bills gambled on Josh having elite throwing power, running ability, size and strength and his intangibles as a smart leader making him able to harness those abilities at a pro-level. 

 

The Bills also gambled that his accuracy issues while there were overstated in college due to the deep passing scheme and low talent level would deflate completion percentage. The Bills so far appear to have been correct, experience and coaching have helped his footwork, decision making and game processing/awareness. Which has tightened up his accuracy and allowed him to harness his elite attributes thus far in 2020. 

 

So I don't blame PFF or FO for not liking Josh out of college. There have been a lot of other "million dollar arm" type QB's who never developed accuracy in the pros. Josh had a good amount of red flags. But after 2019 Josh gave hints of being able to win and improve, you had to at least upgrade your outlook of him somewhat. And thusfar in 2020 he has only shown to progress more and is likely at the very least a mid-level starting caliber QB.

 

The problem is that NONE of these analytical projections of Josh as a pro bothered to look at WHY he had accuracy (read: completion %) issues at Wyoming.  They simply said “he’s inaccurate” and rolled from there.  What I and several others around here have been saying for the past two years is that there were very good reasons why Allen was not a polished product in college — he was light years behind the other QBs drafted in his year in terms of experience, reps, coaching, etc.  But what he did have — freakish athletic ability and desire — were what those who liked him didn’t overlook.  Beane and the Bills scouts deserve so much credit for believing that with the proper coaching and time Josh could become a very good pro.  All we have seen from Josh is a consistent trajectory upwards, and I posted multiple times this summer that there was NO reason to believe that would not continue given his work ethic and the tools at his disposal.

 

The people who are still clinging to their pre-draft analysis of Josh Allen are fools.

 

[Coach Tuesday]

33 minutes ago, eball said:

 

The problem is that NONE of these analytical projections of Josh as a pro bothered to look at WHY he had accuracy (read: completion %) issues at Wyoming.  They simply said “he’s inaccurate” and rolled from there.  What I and several others around here have been saying for the past two years is that there were very good reasons why Allen was not a polished product in college — he was light years behind the other QBs drafted in his year in terms of experience, reps, coaching, etc.  But what he did have — freakish athletic ability and desire — were what those who liked him didn’t overlook.  Beane and the Bills scouts deserve so much credit for believing that with the proper coaching and time Josh could become a very good pro.  All we have seen from Josh is a consistent trajectory upwards, and I posted multiple times this summer that there was NO reason to believe that would not continue given his work ethic and the tools at his disposal.

 

The people who are still clinging to their pre-draft analysis of Josh Allen are fools.

 

 

You are vastly oversimplifying it and it's partially because you're in a 2-0 frenzy this week and need to calm down a bit.  It's perfectly ok to have a measured discussion about football, eball, even with the Bills off to a hot start.

 

You're misunderstanding what FO does.  They didn't simply say "he's inaccurate" and roll from there.  They create (and constantly tweak, perhaps unlike PFF) predictive models based on historical performance.  What they did was say, based on Josh Allen's college comparables, here is what our model predicts - and it wasn't an optimistic prediction.  Predictive models aren't 100% accurate and, often, they need to be tweaked when an outlier like Josh comes along.  I suspect, if Allen develops into a consistent high-performer, FO will acknowledge as much (as they started to this week) and take a hard look at how their model may have missed and what they can do to tweak it.

 

Not everything is black and white.  Life is complex.  There are a bunch of very vocal posters here who HATE nuance more than anything else and it's exhausting.

Edited September 22, 2020 by Coach Tuesday

[eball]

9 minutes ago, Coach Tuesday said:

 What they did was say, based on Josh Allen's college comparables, here is what our model predicts

 

You are proving my point.  Nobody bothered to go beyond that to see if there are circumstances beyond the stats to explain things one way or the other.  Now, admittedly, that process could be exhaustive if they did it for every prospect, but even a cursory look at Allen's background would have revealed a kid who was a "late bloomer" and didn't grow up playing QB.  Who didn't know he had to send tapes to schools to get a look?  This was an easy one.

 

My arguments over the past two years have not been "black and white" -- they've looked for those exact nuances you are discussing.  Your condescending tone attempting to paint me as a simpleton is misplaced.

[Coach Tuesday]

13 minutes ago, eball said:

 

You are proving my point.  Nobody bothered to go beyond that to see if there are circumstances beyond the stats to explain things one way or the other.  Now, admittedly, that process could be exhaustive if they did it for every prospect, but even a cursory look at Allen's background would have revealed a kid who was a "late bloomer" and didn't grow up playing QB.  Who didn't know he had to send tapes to schools to get a look?  This was an easy one.

 

My arguments over the past two years have not been "black and white" -- they've looked for those exact nuances you are discussing.  Your condescending tone attempting to paint me as a simpleton is misplaced.

They didn’t do a unique analysis of Allen, eball.  They ran the entire QB class through their model and wrote an article predicting outcomes based on past performance.  His prospects were ranked very very low.  They acknowledged today that their model may have failed.  If that’s the case, they’ll try to fix it.  I don’t know what your gripe is.  Football Outsiders is an analytics website primarily for sports betting and fantasy football.  They don’t have the bandwidth to go meet with Josh Allen’s pee wee football coach prior to running their predraft article on that year’s QB crop.  if they had factored in that he was a late bloomer with relatively little football experience, the prediction would’ve been even more negative!
 

They are pretty frank about not being perfect - no model is perfect.  It just has to be better than any other model and if you think all of analytics is useless because it’s never perfect, well, I don’t know what to tell you.

Edited September 22, 2020 by Coach Tuesday

[eball]

12 minutes ago, Coach Tuesday said:

They didn’t do a unique analysis of Allen, eball.  They ran the entire QB class through their model and wrote an article predicting outcomes based on past performance.  His prospects were ranked very very low.  They acknowledged today that their model may have failed.  If that’s the case, they’ll try to fix it.  I don’t know what your gripe is.  Football Outsiders is an analytics website primarily for sports betting and fantasy football.  They don’t have the bandwidth to go meet with Josh Allen’s pee wee football coach prior to running their predraft article on that year’s QB crop.  if they had factored in that he was a late bloomer with relatively little football experience, the prediction would’ve been even more negative!
 

They are pretty frank about not being perfect - no model is perfect.  It just has to be better than any other model and if you think all of analytics is useless because it’s never perfect, well, I don’t know what to tell you.

 

I'm not critiquing their model for potentially "missing" on Josh Allen...I'm not even singling out FO.  I think it's great they came out with a statement that recognized their model may have been way off with Josh.  My point, I think, is that there are so many people who don't know a damn thing about analytics who took these pre-draft projections as gospel and keep hanging onto them as if the worst thing in the world is to accept that sometimes people break the mold.  I saw a fantasy football guy on Twitter "critiquing" Josh's game against Miami because he threw a couple of "almost interceptions."  That schitt doesn't happen to any other QB who has a game like that.

 

Analytics are great; what's better is when they're not viewed as the be-all-end-all.  I admit when I'm wrong all the time; just ask my wife.  It's not a weakness.

 

There were signs that Josh didn't "fit" into a typical analytics model.  Most missed them.

[billsfan89]

4 minutes ago, eball said:

 

The problem is that NONE of these analytical projections of Josh as a pro bothered to look at WHY he had accuracy (read: completion %) issues at Wyoming.  They simply said “he’s inaccurate” and rolled from there.  What I and several others around here have been saying for the past two years is that there were very good reasons why Allen was not a polished product in college — he was light years behind the other QBs drafted in his year in terms of experience, reps, coaching, etc.  But what he did have — freakish athletic ability and desire — were what those who liked him didn’t overlook.  Beane and the Bills scouts deserve so much credit for believing that with the proper coaching and time Josh could become a very good pro.  All we have seen from Josh is a consistent trajectory upwards, and I posted multiple times this summer that there was NO reason to believe that would not continue given his work ethic and the tools at his disposal.

 

The people who are still clinging to their pre-draft analysis of Josh Allen are fools.

 

 

Mel Kiper out of all people was banging the drum on that point that the system he played in during college and the receiving talent was deflating his completion percentage. However Josh did have legit concerns about footwork and game processing that were contributing to his inaccuracy. So it wasn't without risk.

[eball]

5 minutes ago, billsfan89 said:

 

Mel Kiper out of all people was banging the drum on that point that the system he played in during college and the receiving talent was deflating his completion percentage. However Josh did have legit concerns about footwork and game processing that were contributing to his inaccuracy. So it wasn't without risk.

 

OF COURSE there was risk.  Nobody is suggesting otherwise.  The point is that for every negative brought up, one could pretty easily see an explanation.  Bad footwork?  Nobody showed him how to do it.  Game processing?  Not enough reps.  He only really played for two seasons at Wyoming.

 

Yes, he had coaches, but nobody can convince me he received good coaching in college.  Anyway, it's all water under the bridge now.  And yes, Kiper can rightfully claim he was on the Josh Allen bandwagon.  It may wind up being the biggest feather in his cap ever.

[SoTier]

1 hour ago, Coach Tuesday said:

 

You are vastly oversimplifying it and it's partially because you're in a 2-0 frenzy this week and need to calm down a bit.  It's perfectly ok to have a measured discussion about football, eball, even with the Bills off to a hot start.

 

You're misunderstanding what FO does.  They didn't simply say "he's inaccurate" and roll from there.  They create (and constantly tweak, perhaps unlike PFF) predictive models based on historical performance.  What they did was say, based on Josh Allen's college comparables, here is what our model predicts - and it wasn't an optimistic prediction.  Predictive models aren't 100% accurate and, often, they need to be tweaked when an outlier like Josh comes along.  I suspect, if Allen develops into a consistent high-performer, FO will acknowledge as much (as they started to this week) and take a hard look at how their model may have missed and what they can do to tweak it.

 

Not everything is black and white.  Life is complex.  There are a bunch of very vocal posters here who HATE nuance more than anything else and it's exhausting.

 

Great post, Coach. 

 

I think the best example of how success in the pros can change how evaluators look at college players is Russell Wilson.  I loved Wilson coming out of college.  I saw him in Wisconsin's season opener and also in the Rose Bowl, and everything about him said "first rounder".  He went in the third round because the only way that Wilson makes 6'2" (the supposed minimum height for NFL QBs at the time) is in 3 1/2 inch platforms.   Wilson's success has forced evaluators to downgrade QB height, resulting in Baker Mayfield (6'1") going #1 overall in 2018 and Kyler Murray (5'10") doing the same the next year.

 

If Jared Goff and Josh Allen go on to have solid or better careers as NFL QBs, it's likely to  open the door for other QBs from smaller programs or even lower divisions to get serious consideration as top draft picks in future drafts. 

Edited September 22, 2020 by SoTier

[Bleeding Bills Blue]

2 hours ago, eball said:

 

The problem is that NONE of these analytical projections of Josh as a pro bothered to look at WHY he had accuracy (read: completion %) issues at Wyoming.  They simply said “he’s inaccurate” and rolled from there.  What I and several others around here have been saying for the past two years is that there were very good reasons why Allen was not a polished product in college — he was light years behind the other QBs drafted in his year in terms of experience, reps, coaching, etc.  But what he did have — freakish athletic ability and desire — were what those who liked him didn’t overlook.  Beane and the Bills scouts deserve so much credit for believing that with the proper coaching and time Josh could become a very good pro.  All we have seen from Josh is a consistent trajectory upwards, and I posted multiple times this summer that there was NO reason to believe that would not continue given his work ethic and the tools at his disposal.

 

The people who are still clinging to their pre-draft analysis of Josh Allen are fools.

 

 

I watched like 1 wyoming game that year.  He threw a few absolute lasers on hitches and out routes.  NFL throws that he made look easy.  The touch issues were definitely there from what i saw though, he never really learned to throw a deep ball without throwing it on a rope.  Trajectory is huge in the NFL, it allows you to create space and windows that you can't see.  He seems to be learning the value of that.

[GoBills808]

1 hour ago, Coach Tuesday said:

They didn’t do a unique analysis of Allen, eball.  They ran the entire QB class through their model and wrote an article predicting outcomes based on past performance.  His prospects were ranked very very low.  They acknowledged today that their model may have failed.  If that’s the case, they’ll try to fix it.  I don’t know what your gripe is.  Football Outsiders is an analytics website primarily for sports betting and fantasy football.  They don’t have the bandwidth to go meet with Josh Allen’s pee wee football coach prior to running their predraft article on that year’s QB crop.  if they had factored in that he was a late bloomer with relatively little football experience, the prediction would’ve been even more negative!
 

They are pretty frank about not being perfect - no model is perfect.  It just has to be better than any other model and if you think all of analytics is useless because it’s never perfect, well, I don’t know what to tell you.

I mean Schatz did call him a ‘parody of an NFL quarterback prospect’ which seems more ad hominem than analytical.

[Coach Tuesday]

4 minutes ago, GoBills808 said:

I mean Schatz did call him a ‘parody of an NFL quarterback prospect’ which seems more ad hominem than analytical.

That’s fine, they are unabashed Patriots fan stat geeks - totally good with calling them out for being mean.  But slamming the entire analytical approach because it didn’t predict success for Allen is way too harsh.

[GoBills808]

2 minutes ago, Coach Tuesday said:

That’s fine, they are unabashed Patriots fan stat geeks - totally good with calling them out for being mean.  But slamming the entire analytical approach because it didn’t predict success for Allen is way too harsh.

I’m a data guy myself but let’s not pretend the argument that analytics is much less predictive in a game as variable as football hasn’t been had for at least several years now.

[Bleeding Bills Blue]

Just now, Coach Tuesday said:

That’s fine, they are unabashed Patriots fan stat geeks - totally good with calling them out for being mean.  But slamming the entire analytical approach because it didn’t predict success for Allen is way too harsh.

 

Yep - I'm gonna roll with my Brett Favre/Steve Mcnair hybrid comparison