The Michael McCorkle "Mac" Jones Thread

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
15,547
Because cherrypicking stats (some of them not particularly meaningful and team based) over a small number of games without regard to team or opponents doesn't do much?
I mean, I didn't think Tannehill was an elite QB last year despite PFF loving him because I understand context?

I really don't get why "really good for a rookie, good NFL starter" isn't enough for people and they have to find a way that he's OMG THE BEST IN THE LEAGUE !!!!!111!!!W!!!!

Edit- and to go into more detail. To me an elite QB is one who can sustain high level performance regardless of supporting cast, game state, etc. over a long period. So Joe Flacco can have a really good season and win a SB.. he isn't elite, Stafford can put up huge numbers... not elite because when his line struggles he fall apart. Rodgers, Brady, elite QBs because it takes a whole lot of dysfunction for them to struggle. Tannehill, not elite because when it comes down to situations where he NEEDS to throw and the defense knows it he struggles with anything but an immaculate pocket. An elite QB is one who can go out there with a subpar line, mediocre or worse weapons, in situations where he has to throw because his defense is bad and execute more than not. I don't think Mac is anywhere near there yet, but most QB's aren't. He's really good at doing the things a QB needs to do to be a positive on a team with a good defense. Since we are a team with a good defense that makes him an excellent fit, and both the team and his stats will benefit.
It's not really cherrypicking so much as pointing out that he's very much improved over the course of the season. It's not like I'm removing three bad games spaced out throughout the season. It's clear he's been playing GREAT - not just great for a rookie, but great for an NFL starter - since week 4.
 

Morgan's Magic Snowplow

Member
SoSH Member
Jul 2, 2006
20,249
Philadelphia
In some ways, it makes me optimistic. The fact that the team was that successful when the running game was shut down and Mac had a sub-par performance is pretty astounding. The screen game was tremendous. Jakobi and Bourne both had huge games with some great plays. The OL did a tremendous job in pass protection We're getting complimentary team performances that are yielding wins, without everything going perfectly.

It honestly makes me think that if they get it all going at the same time in the playoffs they could very realistically win the Super Bowl.
I'm also pretty sure this was the first cold weather football game Mac had ever played in his life. Not making excuses, but its good to get that under his belt and be largely successful because there will be plenty of those games in the future (starting next week). Throwing a football in 35 degree weather is just different.
 

ragnarok725

Member
SoSH Member
Nov 28, 2003
5,545
Somerville MA
* Apparently he got no credit for the great throw to Bourne for the TD, because Bourne needed to make a great play to make the contested catch. I think the throw was good enough that he should get some credit, unless Mac is somehow to blame for the also great coverage.
Yeah he should definitely get credit for that one. I think it was one of his better balls of the day. The defender has his back turned and he gave Bourne a chance.

I actually think the better read on this play is Henry out of the slot who beats his guy going to the corner and is a bigger target. But it worked out, and the throw was pretty good.

View: https://www.youtube.com/watch?v=5R-Z-KgLOHc
 

Soxy

Member
SoSH Member
Dec 1, 2008
6,072
Aside, the fake precision (51.9? You sure Mac wasn't a 54.3 yesterday?) always cracks me up.
This is far and away the funniest part of PFF grades. I’m already used to seeing this sort of nonsense in my day job (100 pt scale wine ratings are a bit of a running joke to many of us in the industry), but adding even more false precision with a decimal is hilarious.
 

Jimbodandy

Member
SoSH Member
Jan 31, 2006
6,468
around the way
Some discussion on these points:

View: https://twitter.com/PFF_Steve/status/1465354930038923270


View: https://twitter.com/PFF_Steve/status/1465356165060120580


I personally prefer PFF - but it has its own quirks like Mac not getting credit for the Bourne contested TD throw
Yeah this is weird shit.

I'm more conservative on the Mac worship than most, but the air yards is precisely why Mac’s game yesterday was one of his better games, not worst games.

Yes, he had a couple of over/underthrows and an INT that was just dropped. He also had some excellent anticipation throws, challenged the defense downfield, and the back corner throw to Bourne was fucking sublime. Even if Bourne missed that catch, it was where only he could get it (also it worked).

We need him to keep challenging defenses downfield, if we're ever going anywhere. Going 21-25 for 185 yards might score well on the test, but that QB usually isn't deep in the playoffs.
 

Van Everyman

Member
SoSH Member
Apr 30, 2009
21,559
Newton
Yeah. These massive variances in advanced stats for Mac’s game yesterday is only showing their limitations.

It’s kind of ridiculous that we’re having a debate over whether the pass to Meyers was lucky or intentional, irrespective of whether it was, you know, complete. Or whether Mac should get credit for the first TD pass to Bourne. I feel like this is a situation where the old Bill Parcell line “you are what your record says you are” is better than trying to divine intent.

Mac obviously missed the pass to Henry. Sometimes I might argue that the receiver wasn’t in the right place, but in that case Mac seem to have plenty of time to adjust even if he wasn’t. So that credibly should be on him. Other than that, he seemed a little less sharp, possibly due to the cold, but not that much. Thus, the VEF (Van Everyman Focus) grades him out at purple (our third highest rating).
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
15,547
I don't get all that. I just saw Mac complete passes at all levels in all parts of the field yesterday. He had a few misses (a couple of bad ones) but also had a bunch of beautiful throws. If yesterday's game wasn't actually "good", I have no idea what's good in the NFL anymore. Dude completed 71% of his passes, and led the team on 8 scoring drives out of 10 possessions, when his running game was TERRIBLE all game long (til the very end). The offense was all on Mac and he delivered, helping them score 36 points.


EDIT: Put it this way: God help the rest of the league if/when Mac Jones actually puts up a "good" performance if 36 points on 71% completions is a "bad" one.
 

Ed Hillel

Wants to be startin somethin
SoSH Member
Dec 12, 2007
30,177
Here
That was not a “dropped INT,” it was a poor decision, but it was just out of the guy’s reach.
 

Eddie Jurak

Go Leafs Go
Lifetime Member
SoSH Member
Dec 12, 2002
34,173
Melrose, MA
I don't get all that. I just saw Mac complete passes at all levels in all parts of the field yesterday. He had a few misses (a couple of bad ones) but also had a bunch of beautiful throws. If yesterday's game wasn't actually "good", I have no idea what's good in the NFL anymore. Dude completed 71% of his passes, and led the team on 8 scoring drives out of 10 possessions, when his running game was TERRIBLE all game long (til the very end). The offense was all on Mac and he delivered, helping them score 36 points.
I think people are overreacting to Mac's PFF score this week. I'm not familiar with PFF methodolgy, but overall on the year, he rates quite highly (4th in the league headed into the game), so it is not an issue of PFF metric/people shitting on Mac. Their rating scale likes him and to the extent that there is subjectivity in it, it usually seems to work in his favor.

As to the bolded, without really knowing the methdology, I will just say the PFF grade passes the eye test for me. Mac completed some big passess, but maybe against not so much in the way of defensive reisstence. He missed some big throws and made some questionable decisions. Moreso, to me, than he usually does.

All in all, not something to worry about.

Someline listing QBs better than Mac suggested that Joe Burrow is better right now. Maybe. Haven't seen enough of Burrow to comment. Look at the stats, and Burrow is doing some things better this year. But last year Burrow was 2-7 as a starter and his numbers from last year seem closer to Mac's from this year than to Burrow's own numbers from this year. QBs get better in their second season! I'm not going to comment on Mac vs Burrow, except to say that Burrow did not have an obviously better rookie year than Mac is having, and Burrow is better in year 2. Mac is doing well this year and there is zero reason to think he won;t improve in his second year, as good QBs tend to do.
 

Soxy

Member
SoSH Member
Dec 1, 2008
6,072
Just for the record, I take no issue with PFF giving Mac a bad grade, good grade, whatever grade. I don't really care about that at all.

My broader issue is that PFF grades are arbitrary and meaningless, because they don't correspond to anything. It's circular logic disguised as scientific analysis.

Just wanted to make sure I was clear on that. ;)
 

pedroia'sboys

lurker
Aug 26, 2007
639
Newington CT
Just for the record, I take no issue with PFF giving Mac a bad grade, good grade, whatever grade. I don't really care about that at all.

My broader issue is that PFF grades are arbitrary and meaningless, because they don't correspond to anything. It's circular logic disguised as scientific analysis.

Just wanted to make sure I was clear on that. ;)
So basically PFF gave him the worst possible rating on their scale. If only Stidham had started the pats would of won by 50.

They have this game rated FAR worse than
Saints
30-51 270 1TD 3 INT
Chargers
18-35 217 0 TD 0 INT
Carolina
12-18 139 1 TD 1 INT

I know I know PFF grades every play individually and final stat lines are often no where near the grades, but this grade comes no where close to passing the smell test. They had Baramore rated in the 50/60s the first 8 weeks.
 
Last edited:

JokersWildJIMED

Blinded by Borges
SoSH Member
Oct 7, 2004
2,322
Yesterday was definitely not Mac’s best game. Great numbers mask a bunch, and failing to convert in the red zone is not good, and although Mac put up numbers he consistently failed to finish drives, hence the bad PFF numbers. He was terrible at the end of the half, not adequately bleeding the play clock and really didn’t get them in reasonable FG range. He personally left points on the board and made several awful decisions that were fortunately not converted into turnovers (near pick 6, near lateral into the ground on quick out (probably should have been challenged), as well as once again failing to recognize a blitzing defensive back. He made a ton of throws and moved the ball up and down the field, and he played better in the fourth quarter.

Real interested for MNF and Eli and Peyton’s take. Mac has been incredible despite some down moments, and I still think the best comparable for Mac is obviously not Brady, but a poor man’s Sheriff.
 

azsoxpatsfan

Does not enjoy the go
SoSH Member
May 23, 2014
2,978
Yesterday was definitely not Mac’s best game. Great numbers mask a bunch, and failing to convert in the red zone is not good, and although Mac put up numbers he consistently failed to finish drives, hence the bad PFF numbers. He was terrible at the end of the half, not adequately bleeding the play clock and really didn’t get them in reasonable FG range. He personally left points on the board and made several awful decisions that were fortunately not converted into turnovers (near pick 6, near lateral into the ground on quick out (probably should have been challenged), as well as once again failing to recognize a blitzing defensive back. He made a ton of throws and moved the ball up and down the field, and he played better in the fourth quarter.

Real interested for MNF and Eli and Peyton’s take. Mac has been incredible despite some down moments, and I still think the best comparable for Mac is obviously not Brady, but a poor man’s Sheriff.
Just a note because I’ve seen people mention this a few times: the clock management at the end of the first half was not Mac’s fault. Bolden running out of bounds resulted in the poor clock management, and Meyers stepping out of bounds on what should have been an easy catch resulted in the long FG attempt. It’s not entirely fair to pin those on Mac
 

DJnVa

Dorito Dawg
SoSH Member
Dec 16, 2010
46,578
For the record, if Mac played every game of the season like he did yesterday he would have a lower PFF grade on the season than Zach Wilson.
 

E5 Yaz

Transcends message boarding
Lifetime Member
SoSH Member
Apr 25, 2002
74,128
Oregon
For the record, if Mac played every game of the season like he did yesterday he would have a lower PFF grade on the season than Zach Wilson.
Good thing PFF grades don't decide playoff teams
 

lexrageorge

Member
SoSH Member
Jul 31, 2007
14,028
And this is the problem with PFF; they make assumptions on who did what for each play based on their own opinion. It's completely arbitrary, and doesn't take into account context, like the play where Myers adjusted his route and Mac threw to him instead of his first option. If PFF is dinging him for the end of first half plays, like the one where Myers stepped out of bounds before Mac even threw the ball, then they are classic garbage-in/garbage-out methodology. I know we're not really supposed to criticize PFF here, but I still say it's fair to call them out when they are obviously incorrect.
 

DJnVa

Dorito Dawg
SoSH Member
Dec 16, 2010
46,578
Wow that’s incredible.
For the record:

Zach's Wilson's PFF Grade: 56.3
Mac Jones PFF Grade yesterday: 51.9

These would be their YTD stats based on Wilson's playing time this season:

Wilson (actual): 113/205, 57.6%, 1313 yards, 4/10 TD/INT
Jones (yesterday's numbers, scaled to Wilson's PT): 161/224, 71.9%, 2170 yards, 14/0

I don't care how much he missed Hunter Henry by yesterday. No one (NO ONE) can look at those numbers and concoct a way to say Wilson has thrown the ball better. Stats have to at least have some correlation to what we see *OR* when we see the stat we look at something a bit differently. There's no way that this works.
 
Last edited:

PedroKsBambino

Well-Known Member
Lifetime Member
SoSH Member
Apr 17, 2003
26,677
For the record:

Zach's Wilson's PFF Grade: 56.3
Mac Jones PFF Grade yesterday: 51.9

These would be their YTD stats based on Wilson's playing time this season:

Wilson (actual): 113/205, 57.6%, 1313 yards, 4/10 TD/INT
Jones (yesterday's numbers, scaled to Wilson's PT): 161/224, 71.9%, 2170 yards, 14/0

I don't care how much he missed Hunter Henry by yesterday. No one (NO ONE) can look at those numbers and concoct a way to say Wilson has thrown the ball better. Stats have to at least have some correlation to what we see *OR*, when we see the stat we look at something a bit differently. There's no way that this works.
that is a pretty compelling indictment of PFF’s ratings of those two
 

snowmanny

Member
SoSH Member
Dec 8, 2005
13,022
To me the Henry miss looked as if he thought Henry was going a bit horizontal and he went more vertical. A bad miss but it looked worse live than on the replay that showed the throw from Mac's perspective.
 

Red Averages

owes you $50
SoSH Member
Apr 20, 2003
6,976
I’ll take a 5300 yard season, with a 34/0 TD to INT ratio and a 17-0 season where you win every game by 23 (yesterdays stats annualized). I mean wtf are we doing here? Brady was 26/34 with 1 INt, 1 TD and 226 yards. You think Bucs fans are worried about his INT, even though they won? Seems like we’re holding Mac to a ridiculous standard. Guy has played 12 games.

The team is on a 6 game winning streak, just blew out the #1 seed, yes with injuries, and we’re being critical about some missed throws. Last week people were mad because BB played conservative while blowing out the Falcons.
 
Last edited:

Seels

Member
SoSH Member
Jul 20, 2005
4,007
NH
Just a reminder that Brady's 2nd or 3rd best season, the unanimous MVP year, he was like 16th in PFF QB rating.

While it serves a purpose it's like looking at political polling 14 months before an election. No stat is perfect, but honestly just raw passer rating does a better job than PFF a lot of the time. EPA > any/a > qbr > passer rating > pff. EPA has him about 10th overall, but 2nd yesterday, and 1st since week 5. I'm not sure how to split passer rating in that way but I'd imagine it's similar.

I think you could reasonably say Mac Jones has been a top 12-15 QB on the league by whatever assessment you want
 

Cellar-Door

Member
SoSH Member
Aug 1, 2006
24,409
I mean, I'm one of the bigger critics of PFF ratings on here, but....
The point of them isn't to tell you what happened. It's to estimate what would happen on average with those same throws with average WR play and average defensive play. So... if Mac made the exact same throws every week, to the same places, PFF is estimating that he usually would not for example get the freakish Bourne YAC, and that the highpoint ball would be incomplete or intercepted more often, and that the bad throws get picked more of the time, etc. etc.

PFF grades aren't supposed to be tied to stats, that's what the stats exist for, so no, Mac wouldn't likely have the same stats if he made those throws every week... we all know that. PFF grades are ripe for criticism, but making idiotic arguments based on a complete lack of understanding as to what they are is silly.
 

Beomoose

Member
SoSH Member
May 28, 2006
19,118
Wherabouts Unknown
To me the Henry miss looked as if he thought Henry was going a bit horizontal and he went more vertical. A bad miss but it looked worse live than on the replay that showed the throw from Mac's perspective.
If so, it's possible that was two guys trying to correct a disconnect they've had a couple times this season. Mac's thrown a couple endzone misses which went deep while Henry went horizontal, not impossible they both went the other direction this time. Hope they get on the same page, because if so there's our red zone TD problem handled.
 

PedroKsBambino

Well-Known Member
Lifetime Member
SoSH Member
Apr 17, 2003
26,677
I mean, I'm one of the bigger critics of PFF ratings on here, but....
The point of them isn't to tell you what happened. It's to estimate what would happen on average with those same throws with average WR play and average defensive play. So... if Mac made the exact same throws every week, to the same places, PFF is estimating that he usually would not for example get the freakish Bourne YAC, and that the highpoint ball would be incomplete or intercepted more often, and that the bad throws get picked more of the time, etc. etc.

PFF grades aren't supposed to be tied to stats, that's what the stats exist for, so no, Mac wouldn't likely have the same stats if he made those throws every week... we all know that. PFF grades are ripe for criticism, but making idiotic arguments based on a complete lack of understanding as to what they are is silly.
I am not sure based on the above that you understood the critique, frankly. I also am not sure about PFF QB ratings and haven't seen a robust statistical analysis, though I have always been dubious because what they are doing is really, really tough to do well on a consistent basis especially as to QBs. I suspect (again, just my own observations) that QB play is MUCH harder for them to do well than most positions as the range of options and the number of variables they face is so high and this (I suspect) means they simply can't do a QB assesment worth much even as they are used a good deal for other positions.

So, to put in the terms you used, if what PFF determines "would have happened" is too far from what actually did happen on a consistent basis (not just one game, a large enough sample to be meaningful) that suggests the way they are rating is not predictive of anything in the actual NFL. And if it is not predictive at all, what is the utility of the ratings they are generating? Typically the challenge here is to understand the true talent level that underlies the stats. What is unclear is whether PFF is able to do so in a better way than the surface stats themselves do.

Remember, and this is critical, PFF doesn't know what would happen to those same throws with average WR and defensive play--this is not like (say) a replacement level hitter's production, which is a defineable statistical level we can compare against. What they are doing is making a subjective assessment of one act (the pass) against two other subjective assessments (the WR play and the defense) and then aggregating those assessments. That's hard to do well, and the question is whether they are doing it well or not.

I do not think anyone was suggesting the fact that their rating and the QB play in a single game vary means they are wrong. But I very much do get the cynicism about whether their ratings are accurate, and the example of Jones' rating and Wilson's stats is a nice way to show the challenge they face.
 

Cellar-Door

Member
SoSH Member
Aug 1, 2006
24,409
I am not sure based on the above that you understood the critique, frankly. I also am not sure about PFF QB ratings and haven't seen a robust statistical analysis, though I have always been dubious because what they are doing is really, really tough to do well on a consistent basis especially as to QBs. I suspect (again, just my own observations) that QB play is MUCH harder for them to do well than most positions as the range of options and the number of variables they face is so high and this (I suspect) means they simply can't do a QB assesment worth much even as they are used a good deal for other positions.

I do not think anyone was suggesting the fact that their rating and the QB play in a single game vary means they are wrong. But I very much do get the cynicism about whether their ratings are accurate, and the example of Jones' rating and Wilson's stats is a nice way to show the challenge they face.
I understand the critique, but people comparing say yesterday's single game score to Wilson's season, then comping a season based on all yesterday's against Wilson's don't understand what the grade is or what it's saying. Jones' grade is low because PFF thinks he made a good number of bad throws yesterday and not many really good ones, that the bad one's were not picked, and in fact caught, and that he got yards and TDs from big YAC and catches they think were on not good throws, means there is a disparity between his grade and the stats. At the same time... let's say he makes every one of those same throws, but neither TD happens because the WR doesn't catch it or run with it, Meyers never makes an adjustment and the long gain goes away, and his worst 2 throws get picked. PFF would have him with the same grade, but the stats would say something like 18/30 220 yards 0 TDs, 2 INT. He made all the same throws. Would we then say... oh yeah he was worse than Zach Wilson's season because it projects to 0 TDs and 24 TDs... of course not.

I don't like PFF"s grades, I think despite them saying they have standardized everything... you can't really do that, it's too complex and you have to make too many assumptions (on routes, defense assignments etc.) but I also recognize what they are aiming to do, and it seems like half this board wants to look at them whenever Mac grades better than the stats would indicate, but when the stats look good but the grade is low they are outraged because OMG<I"D TAKE THOSE STATS EVERY WEEK !!!!

I don't think the Mac/Wilson comp is at all good for 2 reasons...
1. you're comparing a game to a season, which doesn't really work on any level or any sport, you wouldn't compare 1 game of 3pt shooting to another player's season, you wouldn't compare 4 ABs to 300 ABs in baseball.
2. The variability of these things is high, Mac could as I noted, just as easily have left yesterdays game with 0 TDs as with 2, he could just as easily have had 2 INT as 0. If he throws for under 250 with no TDs and 2 INT then nobody would complain that his grade was worse than Wilson's season grade.

That's why it makes more sense to look at season numbers, because it evens out the luck elements somewhat. PFF's system was unimpressed with Mac yesterday, we can debate whether their system perhaps accounts too little credit to the QB on tough catches for example... but arguing that a 1 game disconnect with outcomes is somehow a flawed process isn't really understanding how the process is supposed to work, or honestly how NFL football games work, there is a lot of luck and other player impact on outcomes once the QB throws it.
 

DJnVa

Dorito Dawg
SoSH Member
Dec 16, 2010
46,578
I understand how it works. And I understand SSS and the comparison.

I'm simply saying that if Mac had an entire season of games not where he threw like that (that throw should have been an INT so we are docking him), but where he had that *exact outcome*, his rating would be lower than Wilson's. That's weird. Yeah, maybe there was just something weird about yesterday's game and it's some anomaly. But man, it just looks off. Especially because I think they are releasing numbers before the all-22 is even available.

I'm not saying anyone needs to stop paying for PFF and I will still look at their stuff.
 

Cellar-Door

Member
SoSH Member
Aug 1, 2006
24,409
I understand how it works. And I understand SSS and the comparison.

I'm simply saying that if Mac had an entire season of games not where he threw like that (that throw should have been an INT so we are docking him), but where he had that *exact outcome*, his rating would be lower than Wilson's. That's weird. Yeah, maybe there was just something weird about yesterday's game and it's some anomaly. But man, it just looks off. Especially because I think they are releasing numbers before the all-22 is even available.

I'm not saying anyone needs to stop paying for PFF and I will still look at their stuff.
He would have to make all the exact same throws, AND get the same outcomes... which is highly unlikely, and that's the point. The grade is quantifying the expected outcome, not the actual one.

The odds of those exact throws getting those exact stats a bunch of weeks in a row is basically zero. If he made different throws and got the same stats he'd have a different grade, because the stats and the grade aren't directly related.

I really don't get why people don't get this.

If someone were grading a baseball player on pitch quality and expected outcome, and he threw 60 pitches, and 20 of them were hung curves, but all the hung curves went for 375 foot flyouts... would you say the quality of his pitches was good, or his outcome was good? When you graded the game would you say... he pitched great? Would you say... wow if he pitches like that every start he's gonna be amazing?

I mean hypothetically, you could hang 20 curveballs a game and have them all be flyouts every start, but that's not the expected outcome, so if you're trying to evaluate the process, you'd say... "those hanging curves were bad, regardless of whether they resulted in outs".

I think you could go through the tape and say "yeah I disagree with their evaluation of these throws", I think I probably would, I disagree with them a lot on what they consider turnover worthy throws for example, and I think I disagree a lot on balls that they put as all WR because it's a tough catch in situations where I think you should credit the QB somewhat with making a throw his guy can go get that the defender can't. But the grades SHOULDN'T take into account things like outcome, because that's not what they are quantifying.

For example.. the long Bourne TD... you could run that play 100 times... how many do we think Bourne scores on? I think next-gen put it at about 1% based on past similar plays and player positions. How much weight do you give the outcome there when evaluating Mac's 3 yard throw?
 

slamminsammya

Member
SoSH Member
Jul 31, 2006
5,896
Palo Alto
He would have to make all the exact same throws, AND get the same outcomes... which is highly unlikely, and that's the point. The grade is quantifying the expected outcome, not the actual one.

The odds of those exact throws getting those exact stats a bunch of weeks in a row is basically zero. If he made different throws and got the same stats he'd have a different grade, because the stats and the grade aren't directly related.

I really don't get why people don't get this.

If someone were grading a baseball player on pitch quality and expected outcome, and he threw 60 pitches, and 20 of them were hung curves, but all the hung curves went for 375 foot flyouts... would you say the quality of his pitches was good, or his outcome was good? When you graded the game would you say... he pitched great? Would you say... wow if he pitches like that every start he's gonna be amazing?

I mean hypothetically, you could hang 20 curveballs a game and have them all be flyouts every start, but that's not the expected outcome, so if you're trying to evaluate the process, you'd say... "those hanging curves were bad, regardless of whether they resulted in outs".

I think you could go through the tape and say "yeah I disagree with their evaluation of these throws", I think I probably would, I disagree with them a lot on what they consider turnover worthy throws for example, and I think I disagree a lot on balls that they put as all WR because it's a tough catch in situations where I think you should credit the QB somewhat with making a throw his guy can go get that the defender can't. But the grades SHOULDN'T take into account things like outcome, because that's not what they are quantifying.

For example.. the long Bourne TD... you could run that play 100 times... how many do we think Bourne scores on? I think next-gen put it at about 1% based on past similar plays and player positions. How much weight do you give the outcome there when evaluating Mac's 3 yard throw?
I get you are pointing out the not great logic attacking PFF rather than defending PFF itself, but it must be said that especially in football removing the context of the receivers you are throwing to and the defenders who are on them to achieve some neutral "expected" outcome is doomed to fail precisely because those factors are a crucial component in good decision making from a QB.

Is Tom Brady throwing those jump balls to anyone besides Randy Moss? If he underthrew a few of them and Randy went and got them anyways, does it make it a bad decision or a bad throw? I get what PFF wants to do but it seems like there is very little signal to squeeze out there.

That, and the fuckin 1000 points of precision.
 
For the record, if Mac played every game of the season like he did yesterday he would have a lower PFF grade on the season than Zach Wilson.
For the record:

Zach's Wilson's PFF Grade: 56.3
Mac Jones PFF Grade yesterday: 51.9

These would be their YTD stats based on Wilson's playing time this season:

Wilson (actual): 113/205, 57.6%, 1313 yards, 4/10 TD/INT
Jones (yesterday's numbers, scaled to Wilson's PT): 161/224, 71.9%, 2170 yards, 14/0

I don't care how much he missed Hunter Henry by yesterday. No one (NO ONE) can look at those numbers and concoct a way to say Wilson has thrown the ball better. Stats have to at least have some correlation to what we see *OR* when we see the stat we look at something a bit differently. There's no way that this works.
My understanding is that individual game grades are essentially not scaled the same way as full season grades (so it's common for a guy to have a full season grade that's higher than his game grades would indicate). From the PFF website (their emphasis not mine)....

Season-level grades aren’t simply an average of every game-grade a player compiles over a season, as the season grade credits the entire body of work. An 80.0 game grade is not close to one of the best games of all-time, but 16 games of 80.0 grades will equal an outstanding season of consistency and likely one of the better seasons in a given year. A simpler example is a three-touchdown game from a quarterback. We’ve seen many three-touchdown games but doing so in all 16 games would be one of the best seasons of all-time.
It is entirely possible that a player will have a season grade higher than any individual single-game grade he achieved because playing well for an extended period of time is harder to do than for a short period.

It's one of my biggest issues with PFF grades - that game grades cannot be compared to season grades in the way you're doing here. I understand why they do what they do but I think it makes for a really difficult to understand grading system, especially when most people are only getting snippets of data because the rest is paywalled.
 

Eddie Jurak

Go Leafs Go
Lifetime Member
SoSH Member
Dec 12, 2002
34,173
Melrose, MA
I understand how it works. And I understand SSS and the comparison.

I'm simply saying that if Mac had an entire season of games not where he threw like that (that throw should have been an INT so we are docking him), but where he had that *exact outcome*, his rating would be lower than Wilson's. That's weird. Yeah, maybe there was just something weird about yesterday's game and it's some anomaly. But man, it just looks off. Especially because I think they are releasing numbers before the all-22 is even available.

I'm not saying anyone needs to stop paying for PFF and I will still look at their stuff.
The counterargument would be that Tennessee's pass defense was unimpressive and that contributed to Mac's success. At the same time, Mac missed a few throws, and a couple of throws were more about great receiver play than Mac's own play. I don't think that's completely true, but it is partly true.

Think of Mac as Dante Bichette and the Titans' pass defense as Coors Field circa 1999.

That analogy only strectches so far. There is a lack of transparency as to how PFF gets its ratings, we know that there is potential for error in what they are attempting, and some things the do seem wrong (like penalizing a QB on a play where the coverage is excellent). It's not hard to believe that they missed on the low side last week. It's also not hard to believe that Mac struggled some last week.
 

joe dokes

Member
SoSH Member
Jul 18, 2005
22,510
He would have to make all the exact same throws, AND get the same outcomes... which is highly unlikely, and that's the point. The grade is quantifying the expected outcome, not the actual one.




If someone were grading a baseball player on pitch quality and expected outcome, and he threw 60 pitches, and 20 of them were hung curves, but all the hung curves went for 375 foot flyouts... would you say the quality of his pitches was good, or his outcome was good? When you graded the game would you say... he pitched great? Would you say... wow if he pitches like that every start he's gonna be amazing?

I mean hypothetically, you could hang 20 curveballs a game and have them all be flyouts every start, but that's not the expected outcome, so if you're trying to evaluate the process, you'd say... "those hanging curves were bad, regardless of whether they resulted in outs".

I think you could go through the tape and say "yeah I disagree with their evaluation of these throws", I think I probably would, I disagree with them a lot on what they consider turnover worthy throws
That's the rub for me. I thought the hanging curve analogy was good for what pff is trying to do. Except you can judge whether a curve is any good in total isolation from anything else. I think judging whether a pass "is any good" (except on the obvious margins) contains so many more moving parts as to make the overall number much more suspect.
 

NortheasternPJ

Member
SoSH Member
Nov 16, 2004
16,451
That's the rub for me. I thought the hanging curve analogy was good for what pff is trying to do. Except you can judge whether a curve is any good in total isolation from anything else. I think judging whether a pass "is any good" (except on the obvious margins) contains so many more moving parts as to make the overall number much more suspect.
Don't forget to throw out all the pitches the pitcher makes against the bottom of the lineup hitters who suck as I'd consider that an "expected pitch" so those really don't count. The Patriots run a ton of high % pass plays like screens and as Lazar pointed out and those are "expected throws" and don't count in PFF world. If you throw a high number of "expected throws" and fewer "normal throws" a single mistake are two are magnified in their ratings vs teams who run a more aggressive offense.
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
15,547
I haven't seen all the PFF grades for the Titans game, but it sounds like they're dinging Mac some for his first TD throw to Bourne. I would question that, if that's true.

0:45 mark of this highlight reel

View: https://www.youtube.com/watch?v=agmDF7U1TVg



It's 3rd and goal from the four, a chip shot field goal is nearly a certain make. The last thing you want is a turnover here, so Josh dials up a play to get a receiver in isolation away from the muck of the middle of the field.

The ball is thrown about 35 yards in the air (doing the geometry from where Mac throws it to where it's caught, given the dimensions of a football field), released just as Bourne is starting to make his break. 35 air yards later, it's thrown to a place where only Bourne can get it. The defender maybe could have knocked it away if he got lucky, but his back was turned to the pass, which Mac knew even as he released it. It was thrown to a spot where it could either have been a TD or an incompletion. Virtually zero chance of a turnover. The play, and Mac's throw, was designed so that it would guarantee the Pats points even if the pass fell incomplete.

It was also thrown to a guy who has shown he can make contested catches like that.

So it was the right throw to the right receiver in the right place at the right time. It was thrown 35 yards in the air to a teeny tiny window that offered a reasonable chance at a TD but reduced the chance of a turnover to virtually nil. And thus it virtually guaranteed that they'd walk away with points.

Could the throw have been five inches better? I guess. But it was one hell of a throw, and obviously a great catch, but that was the play design, for Bourne to make that kind of play.

Run that exact play 100 times with a random assortment of NFL QBs throwing it. How many times does the QB throw it better than that? What even would "better than that" look like? Not much different than what actually transpired. I'm not saying it was a *perfect* throw, but it was a damned good one. If PFF somehow dinged Mac for that simply because Bourne had to make a superb catch, well, their grading system needs some work.
 

PedroKsBambino

Well-Known Member
Lifetime Member
SoSH Member
Apr 17, 2003
26,677
I understand the critique, but people comparing say yesterday's single game score to Wilson's season, then comping a season based on all yesterday's against Wilson's don't understand what the grade is or what it's saying. Jones' grade is low because PFF thinks he made a good number of bad throws yesterday and not many really good ones, that the bad one's were not picked, and in fact caught, and that he got yards and TDs from big YAC and catches they think were on not good throws, means there is a disparity between his grade and the stats. At the same time... let's say he makes every one of those same throws, but neither TD happens because the WR doesn't catch it or run with it, Meyers never makes an adjustment and the long gain goes away, and his worst 2 throws get picked. PFF would have him with the same grade, but the stats would say something like 18/30 220 yards 0 TDs, 2 INT. He made all the same throws. Would we then say... oh yeah he was worse than Zach Wilson's season because it projects to 0 TDs and 24 TDs... of course not.

I don't like PFF"s grades, I think despite them saying they have standardized everything... you can't really do that, it's too complex and you have to make too many assumptions (on routes, defense assignments etc.) but I also recognize what they are aiming to do, and it seems like half this board wants to look at them whenever Mac grades better than the stats would indicate, but when the stats look good but the grade is low they are outraged because OMG<I"D TAKE THOSE STATS EVERY WEEK !!!!

I don't think the Mac/Wilson comp is at all good for 2 reasons...
1. you're comparing a game to a season, which doesn't really work on any level or any sport, you wouldn't compare 1 game of 3pt shooting to another player's season, you wouldn't compare 4 ABs to 300 ABs in baseball.
2. The variability of these things is high, Mac could as I noted, just as easily have left yesterdays game with 0 TDs as with 2, he could just as easily have had 2 INT as 0. If he throws for under 250 with no TDs and 2 INT then nobody would complain that his grade was worse than Wilson's season grade.

That's why it makes more sense to look at season numbers, because it evens out the luck elements somewhat. PFF's system was unimpressed with Mac yesterday, we can debate whether their system perhaps accounts too little credit to the QB on tough catches for example... but arguing that a 1 game disconnect with outcomes is somehow a flawed process isn't really understanding how the process is supposed to work, or honestly how NFL football games work, there is a lot of luck and other player impact on outcomes once the QB throws it.
I think you're addressing a set of challenges different from that raised. If we don't trust the methodology and measurement on a single-game level then I'm not confident combining a lot of games fixes anything. Getting a larger sample works when we trust the measurement and are trying to address variance within any individual sample---but if we aren't even sure about what happened that to me is a different problem.

When we look at a player going 0-4, or missing 4 three pointers, we hesitate to call that an assessment of their true talent level becuase it's a very small sample and there is variance in performance in a small sample. So, more data/more games is a good way to address that concern---to smooth out variance between true talent and observed performance in each sample across a larger one. We do not have a measurement question there---we know what a hit is, and we trust that the recording of whether it was a hit or an out is accurate. Ditto a three pointer. We know what happenend and how to categorize it, and the question is what that game tells us about the underlying skill level.

But the challenge being raised about PFF, and in particular about QB assessments, is different. It is is not only a question of whether someone missing four three pointers in a row cannot shoot as a true talent level issue----the problem here is we don't have any confidence that we're measuring 1) whether shots went in and 2) that their feet were past the three point line. More data, to me, likely doesn't address that at all. Imagine a pickup game, where there is no three point line on the court and no statistician. One of the players says post-game "you were off today, what were you 0-10 on threes?" You might have been 0-10, but you also might have been 3-7....we don't have reliable measurement of what occurred, and we don't have a way to validate how the 'three point line' was interpreted in the data. In that context, I'm not sure that aggregating that same person's impressions of a multiple games helps, and it may increase (not decrease) the problem because we have huge error rates each game, and the games are essentially indepedent because the assesment isn't really the same game-to-game, and thus we don't really get a larger sample of reliable data to help us manage variance.

To test whether it aggregates in a useful way (as opposed to a nonsensical way) we might ask what happens if we play out the above example for a bunch of games (or a season, as DjnVA did). When we do that here, it makes no sense---the game-ratings do not correspond to any realistic 'season' or comparison between players. I get that is not what they are arguing they can do, but I am putting forward we should ask the question of what it says about the reliability of their measurement if we can't do so.

That does not make PFF useless--in particular, I think the QB assessment is very challenging and so their ability to measure what is going on elsewhere might well be much better. But I also believe just saying "they aren't trying to do what you suggest" doesnt' actually respond to the critique that "what they claim to be doing doesn't seem to correlate wtih the real world, and that's a problem"

As an aside, I agree Mac had some bad throws Sunday and so if we had just asked the question 'was he great or not' I have little problem saying he was not. But I'm not sure that PFF is a great way to quantify that concern...
 
Last edited:

DJnVa

Dorito Dawg
SoSH Member
Dec 16, 2010
46,578
He would have to make all the exact same throws, AND get the same outcomes... which is highly unlikely, and that's the point. The grade is quantifying the expected outcome, not the actual one.
Again, I understand that. Even if you post "I don't know why people don't understand that" again. I said it seems WEIRD from the 30000 foot level that a game scores that low when we see what the actual results were--I understand process and results are different. Nowhere did I say "blow up PFF" or "if you subscribe to it you're a moron".

I'd say the same if there was a PFF style pitch grading system in baseball where Eovaldi throws 9 straight meatballs that are swung at and missed and grades lower than Gerrit Cole throwing the world's greatest slider and Devers hits it 450 feet. I 100% understand that the tool would say Cole threw a better pitch.
 

Eddie Jurak

Go Leafs Go
Lifetime Member
SoSH Member
Dec 12, 2002
34,173
Melrose, MA
Don't forget to throw out all the pitches the pitcher makes against the bottom of the lineup hitters who suck as I'd consider that an "expected pitch" so those really don't count. The Patriots run a ton of high % pass plays like screens and as Lazar pointed out and those are "expected throws" and don't count in PFF world. If you throw a high number of "expected throws" and fewer "normal throws" a single mistake are two are magnified in their ratings vs teams who run a more aggressive offense.
I realize we are all debating the application of a non-transparent methodology, which is itself problematic. But. The implication of what you suggest here would be that a QB like Mac won't rate will in this system, which I think is false - the Mac has been rated highly by PFF thorughout the season until this one game.
I haven't seen all the PFF grades for the Titans game, but it sounds like they're dinging Mac some for his first TD throw to Bourne. I would question that, if that's true.
I think it is a clear limitation of their system. Mac threw a great ball, but the coverage was also great. I think on most throws, that ball with that coverage is an incompletion. On the other hand, that ball with less great coverage is an easy TD on most throws. Mac had to release early. Was there something he should have seen before he threw it that should have led to him not making the throw? Was Mac supposed to look the defender off first? I don't know, but even if there was some way he should have handled it differently I think it is in the realm of parsing things too finely. I'm guessing it is a general problem of the system that when a QB does something well but the defense also defends it well, the QB gets dinged more than he deserves.

However, I think any QB rating system is going to run into problems. I would guess that PFF did not ding Mac for Meyers' error of setting up with his feet on the line, a mistake that led to Mac's pass being ruled an incompleteion and that probably cost the Pats an FG. A more objective stats based system probably dings Mac for failing to complete an easy pass.
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
15,547
Also, we have no idea who each pass play is really designed for. Say Jones drops back and has a small window available for a chunk play downfield, but there’s risk involved. But there’s pressure so he decides to take the short receiver for a 7 yard gain to set up 2nd and 3.

Was that a good play by Mac or not? Who knows. Mac seems to take the safe route a lot and there was a bunch of criticism that he wasn’t taking shots. So ok he takes deep shots and it’s hard to put the ball EXACTLY where it needs to go. So on the deep contested pass to Meyers he takes a shot and it’s complete but he gets dinged because it could/should have been deflected away?

I just know that the guy, on a day when they had NO running game til the last few minutes of the game, marched them up and down the field and got points on 8 out of 10 drives and they put up 36 points as he completed 71% of his passes with zero turnovers.

If somehow that’s not a good game, then the rest of the NFL should be quaking in their boots because when he actually DOES have a good game, it’s going to be unreal.