Jimmystats: Projecting Patterson and Peters
How up is up? [Bryan Fuller/Patrick Barron]
In the last spring football bits, and then in the podcast, I mentioned that I’d pulled the sophomore/redshirt freshman stats of a couple decades of quarterbacks we know in order to put our two 2016 signal-callers’ 2017s in perspective. I figure I should show that work.
I was trying to answer two questions: where did Peters/Patterson rank among other QBs their age, and what’s a standard progression for a guy who started his second year upon entering his third?
1. Shea and Brandon vs Other 2nd year QBs
Bless Foxsports.com for keeping sack stats with quarterbacks. I wish I’d taken rushing stats too but alas I only grabbed the passing stuff.
There’s going to be a lot of noise in here so a few things as you peruse:
- Some of the weirder results were low sample sizes. I only counted 100 or more attempts, so like Tate Forcier 2010 didn’t make the cut and Peters just did barely. Still I shaded in red a few to beware.
- I have them ranked by “Efficiency Percentile,” a junk stat I made up that’s 49% sack-adjusted yards per attempt, 23% each touchdown and interception rate, and 5% completion rating, normalized as a percentile among all the QB seasons (about 500 of them) I grabbed. The QBs are Big Ten mostly plus ND and Stanford guys Harbaugh coached. Kenpom Christmas colors apply: green is good, red is bad.
- I also showed completion %, sack rate, and college quarterback ratings, though they’re not included in the rankings.
The Kizer comp for Patterson is pretty close—I’m hesitant to give him Russell Wilson because Wilson had an extremely low interception rate, a trait which seems to presage NFL success. Shea…does not. The good news is the sack rate is correlated with so many things it’s impossible to say if Shea’s higher than normal rate of going down behind the line was indicative of anything. Sometimes it’s just a bad OL.
It was very possible for two QBs on the same team to have dramatically different sack rates. But the stat doesn’t give the reason. Sometimes that really was about poise:
Sometimes however it’s about which part of the schedule you got:
And sometimes it’s completely counter-intuitive:
Other than Denard most running QBs had high sack rates like Shea’s. I can’t tell you why—maybe those guys try escaping instead of throwing it away when under pressure, or maybe being part of the run game gets them sacked because of play-action. The highest sack rating I tracked was Braxton Miller in 2011. Two of the top four are notoriously stoic T.J. Ostrander seasons, and two more in the top ten are 2007 Notre Dame (bad OL). It’s too much to unpack.
[Hit THE JUMP for the Peters projection]
-----------------------------
As for Peters—and warning: low sample—he was not at the bottom but certainly in the mediocre range of guys his age.
T.C. Ostrander in 2004 (two years before Harbaugh) is the nearest comp except for that afore mentioned sack rate, since both were averse to interceptions. Tanner Lee’s year at Tulane, and Danny Etling at Purdue were a few other similar players at his age. However neither of those guys were as efficient as our guy. Expanding that to any class here are some quarterback seasons similar to last year with Peters:
Some duds in there but also plenty of “don’t give up on that guy!” Brian Hoyer had an NFL career after that. The reaction most got was “eh, fine,” unless it was coming from Penn State fans.
We might as well swing back to Patterson then.
You’d take any of those seasons.
----------------------------
Question 2: What’s a Standard Year 2-3 Progression?
My sample wasn’t really large enough to get good read on this because I would need a much larger sample. There were only 39 quarterbacks I pulled data for who had 100+ attempts in both their 2nd and 3rd season. A median improvement in my efficiency stat was +5%, but that could vary from +22% (Carlyle Holiday) to –17.7% (JT Barrett).
bliggering it makes it clicker
Red means they got worse, green means they got that much better. Shea would be starting between Kizer and Russell Wilson and peaking over the 70% line. Peters would starting next to Juice Williams, and a standard progression should see him get over the 50% line.
To make my projections I used the median percent change in each stat from this tiny sample and applied that to the sophomore/RS freshman seasons of Patterson and Peters. This is fraught with problems, for example INT rates generally went up with a rise in attempts and YPA. Is that a bad sample thing or is it 3rd year quarterbacks trying to do more?
Anyway the very salt-requiring standard progression should put Patterson just outside the top 25 best QB seasons in my study and Peters a bit below the middle.
Projected Shea would be in pretty strong company:
Projected Peters not so much:
I couldn’t include Brian Lewerke because of his freshman sample size, but he was another good comp for Peters. These guys don’t suck; they’re slightly below average starting quarterbacks. Patterson’s projection is more like “pretty good,” and getting into all-conference in a down year.
If you want the whole list it’s here:
April 25th, 2018 at 12:54 PM ^
The 'QB Efficiency Deltas' figure is a beautiful piece of data visualization.
April 25th, 2018 at 12:56 PM ^
Patterson's o-line was also bad, not a good excuse for Peters.
He also had freshman receivers which compounded that fact.
Hard to be a pro style QB and stay in the pocket when you have no OL. Patterson may have also had a bad OL, but Patterson has wheels and can get out of the pocket a lot better than Peters.
April 25th, 2018 at 12:59 PM ^
The QB Efficiency Delta's is a fantastic visualization! And super encouraging!
Probably it's completely misleading though. I'd guess the big reason that quarterbacks improve is that they become more comfortable in the system. That will apply to Peters and not to Patterson.
who will win the QB competition. My guess it will be Patterson but if Peters or Dylan plays well either of those two may win the QB race. I'm just glad we have 3 talented young QBs battling it out. Our QB position should be much improved this season. I just hope our OL shows improvement this season.
Danny Etling: 10-3 LSU team
Bryan Cupito: 7-5 Minnesota team
Troy Smith: 12-1 OSU (Lost National Championship game to FSU)
JT Barrett: 11-2 OSU team
Ricki Stanzi: 8-5 Iowa team
That's a healthy mix of good and mediocre results to go along with those statistical QB rankings. I was hoping that those QBs would have all lead strong outfits as well (though I recognize that wouldn't have been a true correlation). Not the case. Just goes to show the true team nature of football.
Hopefully either one of these guys is good enough to help the team be successful.
First, I just want to say thank you for Seth for putting all this together--awesome job man. I don't think some people realize how long this can take to pull togther & it's like mana from heavan in the off season for a college football junkie.
Second--Guys--cut Seth some slack when he's on the MGoPodcast. It seems like you were running him through the ringer there a bit. Nobody puts Seth in the Corner!
Third--and to your point RE: various season records being a bit over the map with the QB years just above Shea's hypothetical 2018 season---all true observations, but obviously context matters. Like the strength of that team's defense overall:
Defensive S&P Ranking for the following seassons/teams:
- 2017 LSU--D S&P #18; Record of 9-4 (Etling)
- 2005 Minnesota--D S&P #71; Record of 7-5 (Cupito)
- 2006 Ohio State-- D S&P #1; Record of 12-1 (Smith)
- 2017 Ohio State--D S&P #8; Record of 11-2 (Barret)
- 2010 Iowa--D S&P #19; Record of 8-5 (Stanzi)
Tons of caveats apply of course, but to me it seems like when you look at Seth's made up "QB Efficiency Percentile" stat & the grouping right above "Shea 2018" line, what you get is possibly a pretty decent correlation (admittedly not causation necessarily).
I'm not a fancy stat guy, but it seems to me one observation could be, "If a team's starting QB gets a QB Efficiency Percentile rougly in the projected Patterson 2018 zone, you can rougly expect the season record to break down along the lines of the team's S&P Overall Defensive ranking."
In other words, if a QB can get to that level, then that year the offense overall won't be a net drag/hinderance on the defense. Doesn't mean they're necessarily good enough to pull UP or compensate for a bad defense, but the team's season record won't significantly under-perform their Overall Defensive S&P ranking.
Michigan's Overall Defensive S & P since Coach Brown has arrived:
- 2016--#2
- 2017--#10 (with a ton of new starters)
- 2018--????? (with only 1-2 new starters & losing almost nobody from the 2nd string)
So, my totally unscientific feelings-ball conclusion just eyeballing Seth's QB Efficiency Percentile stat versus Overall Defensive S&P ratings is this:
Assuming this 2018 season:
- Michigan Defense can stay in the top 10 S&P overall l ike we did in 2016 & 2017 (I think a pretty reasonable assumption); and
- IF Shea gets eligible, avoids major injury, and he performas at or near the "2018 Patterson Projection" line that Seth cooked up (lots of IF's there....); then...
.....it's not an unreasonable projection to say that what we're looking at is a 10+ win season.
Again, major caveats apply (strength of schedule, youth of the rest of the roster, avoiding key injuries, flukish football-god type events not breaking against our favor disproportionately). And some optimists have been saying 10+ win season is doable since the saltiness of our bowl game has worn off. But it seems to me that if you pair Seth's post & the Defensive S&P observation, there's a least a bit more hard data that seems to point to that 10+ mark as being a real possibility beyond mere hope/feelingsball.
Nice job with this dive into the numbers. They would seem to justify fan enthusiasm for Shea.
I don't think Peters had enough attempts to conclude much from his numbers, however. 108 attempts is not very many, and 44 were in the bowl game.
i may be way off base here, but the guy is a really solid college quarterback behind an OK line. Definitely seems above average and could compete to be the best QB in the big ten this year
He is good dont get me wrong but Cook and Cousins were way better than him
As a freshman Lewerke was in that same range as Peters: meh. Like he's not a total disaster (you'll see a few of those lower down) but his job was to not screw up, not lead the offense.
Lewerke last year progressed about an average amount and was effecive. That's better than meh, it's still not dramatically adding to a team's likely wins. That's where I would expect Peters to be next year based on this study.
I want to look into deltas more and see if some stats correlate higher. My guess is low INT rates end up better QBs.
Yesterday I played around with 3rd to 4th year deltas and found a weird chart: the good QBs almost all DROPPED, and the bad QBs nearly all rose. On inspection a lot of that were redshirt juniors losing their jobs to better young players, and a lot of the good QBs who might have risen were early NFL entries, if not always for good reasons (Kizer hated his coach, Hackenberg was about to lose his job, Pryor was suspended, etc.)
If he played under rich rods offense he wouldve been amazing
April 25th, 2018 at 10:57 PM ^
Brandon Peters is probably better than his comparables make him seem, just because he didn't play the non-conference part of the schedule.
For example, If you drop Patterson's nonconference part of the schedule, his completion percentage drops to 58% and his QBR to 124. His interception rate is 3.7% and his sack rate is 8%. Suddenly he would look like he falls somewhere in the Tommy Armstrong -- Alex Hornibrook range. (Just to be clear -- he's not in this range. Armstrong and Hornibrook also had non-conference opponents in their stats.) That's about an 18 to 20 spot drop in the rankings, and would leave him looking just a little bit better than Peters.
So it's fair to assume that had Peters gotten the benefit of playing some non power five schools (and even non FBS schools) like Patterson then he'd look 18 to 20 spots better, and we'd be thinking "Better than Chad Henne, worse than Braxton Miller".
Peters got to play Minnesota, Rutgers, and Maryland who combined to win 7 conference games and where Michigan rushed for a combined 875 yards. That's pretty non-conference like right there.
And Michigan could have given him some garbage time reps in those games to pad stats and improve. They chose not to.
About 60% of Peters' attempts came against Wisconsin and South Carolina. The worst team (by far) that he played was Rutgers. And as much as we like to make fun of Rutgers, they'd (probably) be an above average MAC team. Patterson's statistics include FCS competition.
I'm not saying this to take anything away from Patterson. Just that Peters is better than the comparisons make him seem.
This is an excellent point, but unforunatley probably impossible to break down in any meaningful way.
But that does largely remind me of the whole Tom Brady vs. Drew Henson QB battle back in the late 1990's.
Henson was always the better raw athlete, so his skillset & ceiling would theoretically show up earlier in his development. He never played much in the league, but it's clear the uber athletes at QB also tend to see a drop in their careers earlier as their legs start to give out (not unlike RBs who get their bodies beaten up). Conversely, Brady's major skillset is the field vision, football IQ, killer instinct/psychological advantage. That stuff usually doesn't bear fruit until a few years of development, but the up side is that it also enables a QB to play longer in the league.
At the college level--especially with the newer 3 & out model--it seems like a college team may be better off with that raw athlete b/c they'd be more likely to get an extra season or so of productivity out of him, whereas going the cerebral route may give you a great upperclassman QB for a year or so, but you may only have a 1-2 year starter on your hands versus a 2-3 year starter.
Double post, don't know how to delete
You've been on here for TEN YEARS and don't know how to delete a double post!!??
Kidding
Totally agree, but it's important to remember that 2015 was a lost recruiting year. Harbaugh had 3 weeks to look around the recruiting landscape & grab JOK out of desperation and whiffed. Then he saw what he had in the QB room in the spring of 2015, grabed Rudock & turned him from a soon to be medical student into an NFL propsect within a season.
And as others have said in this thread, Peters--our 2016 QB recruit-- had a super awkward 2017 development path. He was bound to make freshman type mistakes, and once the competition level ratcheted up at Wisconsin that required him to perform beyond the simplistic game plans against Rutgers/Minnesota/Maryland, those showed. The bowl game was an offensive shit-show, but I'm not ready to put a ton of blame on Peters for that as in my view, in terms of snap counts, that was basically his 4th game of his entire career. If he keeps making those errors into Big 10 season without signs of imporvement, then I'd worry about Peters, but not before then.
Bottom line is we've got a ton of bullets in the chamber at QB now. So, one of two things will be true one year from now, regardless of whether Shea was eligible for the 2018 season, whether Shea did or didn't do well in 2018 if he does play, whether Shea enters the 2019 NFL Draft, or whether he decides to come back for the 2019 season.
Either.....
- We will have ZERO concern as to the QB position going into the 2019 season (some combo of Shea returning in 2019; Peters improving to the point of being solid; McCaffrey and/or Milton being very solid);
OR
- Harbaugh will be on a very hot seat because it's clear we don't have a QB solution on the horizon & the #1 crtiticism of him will be--irony of all ironies---his utter failure to develop a reliable QB position for any sustainable period of time.
I think the likelyhood of the first outcome is 95%+ chance and the #2 outcome is exceedingly unlikely to happen.
But that's entirely the point of hiring an "elite" coach. To make the players better. No it doesn't mean that JH can make guys into NFL Tom Brady overnight, or at all....but given his reputation for offense and QBs, our QB situation should NOT have been this poor for this long. There has to be some consistent improvement at some point. And yet Speight totally regressed, JOK completed melted, and Peters often looked like a deer in headlights. These are all guys who had been "in the system" for at least 2 years (Peters was a redshirt freshman, not a true freshman, so it was his second year in the program) and yet NONE of them could even produce an average to slightly above average performance on the field?? There's a problem with that when you're in the middle of a coach's 3rd season who's supposed to be a guru.
JH needs to step his coaching and development game up on the offensive side of the ball....bottom line
Thanks Seth for keeping football alive!
EDIT: Apologies for formatting. Don't know how to fix some of it.
Slow day at work so apologies for the longer posts/replies, but I've got nothing better to do today, so...
TL:DR Version--no we are not a basketball school now, I dont' think Brian quit but my guess is he is doing a reorganizaiton behind the scenes so that other folks will be producing more of the football content from now on--which actually is a good thing for everybody involved.
I obviously have no special knowledge of any of this since I'm just a random poster, but here's my read on the situation--
We aren't a basketball school & I sincerely doubt that the blog content will move AWAY from football. Maybe as they grow, they grow the basketball/hockey content MORE so that as a percentage, the football content shrinks. But I don't think that the actual raw amount of posts/content for football has or will drop in any meaningful way.
If I had to guess what was really going on w/ Brian, it's a mixture of this:
- He's a grown ass man with a young family and doesn't want to be a shitty husband/father. His kid I think is a toddler now? If so, then even more than infants, the toddler phase requires just a TON of time & attention. Infants distrub your sleep patterns, but you can just put them in a rocker as they cat-nap all day long & you can blog your ass off for hours at a time mostly uninterrupted. Not so with toddlers. They require actual human socialization and active monitoring.
- Football may just be the easiest content to delegate. He's got a solid staff that has a pretty damned good handle on the football stuff already. It's a more popular sport, so the ability to find new/additonal staffers to crank out the football content in the future may be easier than some of the lesser sports. Lastly, I think the annual/seasonal football features/content is older & thus perhaps more "mature" in the development phase--more templates, samples of older work, etc that can be used to coach up newer staffers into how to produce good quality football content in the future.
- As was the topic this winter, if he is broken/burned out by the football stuff, I don't think it's in a way that some casual observers might think. I think a lot of people in the heat of the moment (myself included) misinterpreted his infamous "Whatever" post after the bowl game as totally giving up on Michigan Football writ-large. I don't think this is really the case. The way I understand it now is that it isn't that Brian PERSONALLY is giving up caring about Michigan Football. And it isn't that he thinks the program is totally doomed and thus not worthy of his attention or passion/heart or antyhing. It's that he's frustrated with some of the meta issues that come along with directly interfacing with some of the crazier HOOOOOTTT TAAAAKKKKE aspects of our fan base who swing radically up or down disprportionately to the most recent event on the field.
At first when I realized this may be the case, I was offended. After all, I am a loyal reader of this blog. And even though I'm not one of the big Hot Take perpetrators, I certainly don't feel good being lumped into a guilt by association thing. Plus feeling like our vaunted leader might disdain some of his own readership offended my sensitbilites. One ought not bite the hands that fees you after all.
HOWEVER, now that the bitter disappointment of last season has faded, I don't feel that way at all.
First, as a practical matter, this is Brian's own blog. It's his house & he can do & say whatever he damned well pleases. We aren't even paying customers. Second, upon reflection, I think Brian stepping back (if he so chooses) from some of the public-facing aspects of the football content is actually an indicator of several positive things under neath the surface.
In no particular order:
- Assuming my interpretation of the true meaning of the "Whatever" post is somewhat accurate, it shows that he actually DOES still care and have a passion for the fortunes of the football team. If he was truly detached & didn't give a rip about the squad, then the comments & hot takes from his readership about the team wouldn't affect him. And yet it does.
- Assuming he does in fact still care about the footbal team & therefore experiences amplified highs & lows because he has to deal with some of the crazier aspects of our fan base to boot, then taking a step back is a healthy/mature level of self-awareness on his part. Viewed through that prism, it is a move towards self-preservation. This means that in theory, although Brian himself would be less directly involved in the football content, it's also a more emotinally sustaining model for him in the long run. So less direct content from Brian, but the fewer times he chimes in directly it'll be on his own terms & theoretically it would enable him to either expand the non-football related content of the blog or alternatively give him the stamina/interest to maintain the blog overall for a longer period of time (versus going on an emotional bender after emotional bender & then burning out in a much shorter time).
- Related to the above, on a logistical level it also means delegating content to his staff members. This again means more blog sustainability over the longer term. If Brian takes long hiatus or something for personal or financial reasons, it isn't like the the blog has to come to a grinding halt. He's got a team that can continue on if Brian (or any other staffer for that matter--Ace & his health issues as a recent example) need to take a step back for any reason.
Anyway, that's my super long totally uninformed guess as to what may be going on w/ the blog behind the scenes.
I think last season required a few months of "nah" especially with the lame recruiting finish and Brian was as much a victim of that as any fans.
We were all burned out with the two final four runs. I didn't have to do as much hoops and hockey coverage so I picked up football.
Staff-wise keep in mind it's only three of us who do this full-time: Brian, Ace and myself, and guys with real jobs who write for us are going to necessarily have to work around their primary incomes. So when we move stuff around it's really just between three people. Next year I think we want Ace to do more basketball, which means I'll be picking up some of his football duties.
The worrying thing about Peters is a low completion percentage despite an extremely low YPA number. The thing that worries me the most is the fact that the staff clearly didn't trust him with the full playbook despite being in the program for two years. Okorn might not have gotten the job done but the coaches clearly trusted him with more of the playbook despite having the same number of years in the program. Peters has the physical tools but still seemed a step slow on the mental side last year. Of course that can change for a young player and it can click, but until it does in a live game situation its a wait and see approach.
Patterson really just needs to give us slighly above average play and we win at least 10 games.
Right, because regardless of the actual on-field results, the state of our football program is always sunny and amazing!
How about you stop being a jerk to people who hold a different, and slightly more realistic/down-to-earth opinion than you?
The reason optimism would take a hit is because Peters and everyone else on the offense looked so bad last year, that there's zero FACTUAL reasons to be inspired and optimistic about this upcoming season.
Yes we're all HOPEFUL that things will turn around, massively, but until that ACTUALLY happens, then it's just that....hopeful and wishful thinking.
Regardless of who our coach has been, we hear the same spring ball and pre-season hype year after year...and yet the on-field results usually look drastically different from the hype. One of the reasons we get picked on by so many other school fan bases.
So excuse some of us if we are waiting to actually see a good team on the field before we start singing the praises of 2018's UM championship team.
but if Peters is the QB our offense will be only slightly better - my view is going 8-5 or maybe 9-4 is highly likely. The 2018 schedule is brutal. However, if Shea is the QB all season and produces those QB efficiency numbers ? ... B1G title is possible. Again as many have said on the board, our defense is elite and maybe top 5, which will keep us in virtually every game. Shea's playmaking ability is the key difference on the road IMO versus ND, MSU and OSU. If Peter's is the QB, I don't see enough offensive production to be effective in those road battles. Shea's eligibility and season-long health is key to being a contender for the B1G.
The level of detail in your analysis is off the charts Seth. Looked forward to this read and it didnt disappoint. Now to continue to monitor Twitter hoping for good news on Shea
Milt Plum.
Thank you. Finally someone just comes out and says it. If Michigan had played 2017 with even slightly below average QB play, there's 2-3 wins to go with an 11-2 record. M fans should be praying that the QB is Patterson/McCaffrey in 2018, with no other QB's stepping on the field.
Comments