Pro Football Focus individual grades for Patriots roster

[RobertWeathers]

Perhaps the players are graded vs others across the league?

In a down NT year Shelton is not so bad.

I thought early in the year he was getting shoved around but he has shown improvement. Hopefully he can sustain it.

 

[maineman209]

so “above average” is not as good as “good”?when i was in skool, dey were teh same

It's kinda like the Celsius-Fahrenheit thing. For some people, "average" = " bad" in the same way that 32F = "effing zero!" Celsius.

Ergo, "above average" = "mediocre", and "good" = "average".

 

[bigballas]

I quit reading after D Mcourty was not graded "Elite"

now it makes sense why BB took him out of that last play in MIA

 

[TB_Helmet]

The NFL’s Analytics Revolution Has Arrived

I posted this a few weeks ago but worth reposting. Just shows that NFL teams are increasingly using analytics largely based on player tracking on the field. My guess is that it’s better than PFF but what do I know. I’d kill to get my hands on that data...

 

[TinhWoodhead]

thats really true!

 

[primetime]

Hold on. So, you were ok with PFF, but not anymore? I'm confused; am I misinterpreting or are you saying you were ok with them grading on a -2 to +2 scale? Or is the "basic number counting" you are referring to something totally different?

Anyway, they still start with the same -2 to +2 scale:
From there, the grades are normalized to better account for game situation; this ranges from where a player lined up to the dropback depth of the quarterback or the length of time he had the ball in his hand and everything in between. They are finally converted to a 0-100 scale and appear in our Player Grades tool.
PFF Player Grades | Pro Football Focus

I'd want PFF to find a way to identify the Xs and Os & tendencies in playcalling. If you are going evaluate every player on every play why not provide some info on what passing concepts are used most? Same for defensive coverage. I don't expect the exact play-call on every play, or anything impossible. But at least identify recognizable passing plays when it's there (i.e. "Spider 2 Y Banana") & what the Defense did (Cover 1); even if you can't tell, just list middle of the field open vs closed, man vs zone, or just an individual route, etc. Don't know why PFF won't do this. It would be much easier to watch the All-22 with some info.



If there's "hype" surrounding PFF it's due to the immediate backlash it gets from fans over its evaluations/tweets. So, if there's hype; its more like controversy, because any time someone cites PFF in a discussion, any discussion, it's immediately met with the same outright dismissal (& valid criticisms) that we are currently discussing.

Also, not "worth a damn"? Really? I totally acknowledge the gaps/flaws of PFF, but you can't just dismiss having former NFL Scouts grade each play of each game. Especially so, because you cannot use conventional stats to evaluate, say, a LT or an ILB like you can with, say, a QB or RB. So, if you want to try and rank all the players at each position; PFF is pretty much the only site that even attempts to do this. Now, back when it was some random fan (from the UK watching a TV broadcast)? Yeah, it was indeed not "worth a damn." But when former NFL Scouts grade players after watching All-22 film...then I genuinely think it shouldn't be outright dismissed. Questions/Criticism of the results? Sure. But not dismissal.

I agree that having trained eyes watch a lot of football is good. The problem is PFF removes all the richness of qualitative analysis in order to tap into the mystique of numbers and mathematical analytics, which has the unfortunate effect of collapsing a lot of the analytical nuance that PFF's scouts could talk about into near-meaninglessness.

PFF wants you to believe that a 90 grade is a 90 grade, but what does that even mean? There's simply no way to parse or present in a meaningful way how a 89.6 differs from a 90.2 in their grading system. This isn't statistics, it's just crap marketed to look like statistics.

 

[luuked]

This isn't statistics, it's just crap marketed to look like statistics.

Bingo. It is quantified, not reproducible gut feeling.

 

[primetime]

its like when i drove for Uber, some riders don’t know that a 4-star rating is actually not good

This is of course due to the fact that internet star ratings are really just a representation of how many 5s versus 1s something gets, especially since the people who rate something are generally those who had either notably good or bad experiences (or are kooks, which is not mutually exclusive, or are bots or people whose ratings are bought by the company).

Very few people provide a 2, 3, or 4 star rating. Like with PFF, there's a certain sort of mysticism here. It would actually be more useful for the consumer if ratings like this were purely binary. Tell us the percentage of ratings that were Good as opposed to Bad. The 5-star system is doing that (since the vast majority of ratings are 5s or 1s) but it's all mystified in an attempt to allow more nuance, so a 4-star rating is someone with a wildly above average number of 1-star ratings.

 

[PP2]

Obviously not in terms of "long speed". In terms of quickness on CoD (where "a step" is much more critical), he may have.

However, it may only have been critical earlier in the season when DMAC ended up in certain less-than-optimal coverage situations (he's not, strictly-speaking, playing "safety"). As JMAC has improved and Jackson has come along, he seems to been much more consistently effective.

Yes, I understand yours and @robertweathers 's point- but as a deep safety, that twitch isn't quite so important as it would be if he were a CB on the other team's #1 or #2 WR, so to speak.

 

[patfanken]

It's PFF, so who cares? PFF grades and analytics aren't to be taken seriously.

As its been noted, the PFF that started as fan driven focus group years ago, isn't what it is now.

I think it can be taken seriously, BUT also with a grain of salt since they don't know the calls or how a specific look has been coached. AND of course, there is the matter of subjectivity that comes with giving a grade on every play.

So PFF is just a good metric for giving us an INDICATION of what a player is doing, based on the limited information at hand. It's data that we can use or ignore as we see fit. But at this point, DI, it would be wrong to say there analytics shouldn't be taken seriously. Is there another organization that does it better, or are all the analytic sites simply BS by definition.

If there are other groups that do it better, maybe we should look at THEIR grades and compare them.

In the end the only grades that count are the ones the staff gives out on a weekly basis. They are the only ones that have the requisite information, and even then subjectivity plays a role, as it was noted.

BTW- I thought they did a decent overall job. They essentially confirmed what we already knew. The Pats are made up of few stars and a lot of better than average players. We should also remember that the game of football is the one game where the sum of the individual players CAN be a lot better than the individual parts. The best talent doesn't always win.

 

[primetime]

The NFL’s Analytics Revolution Has Arrived

I posted this a few weeks ago but worth reposting. Just shows that NFL teams are increasingly using analytics largely based on player tracking on the field. My guess is that it’s better than PFF but what do I know. I’d kill to get my hands on that data...

NFL Big Data Bowl | NFL Football Operations

Have fun. I suspect, as with most data of this sort, there's way too much noise to test any useful hypotheses (and I think a lot of this big data sports analytics stuff is wasted time and mistaken inference, and I'm a statistician by trade!), but it'll be cool to see what people do.

 

[maineman209]

Yes, I understand yours and @robertweathers 's point- but as a deep safety, that twitch isn't quite so important as it would be if he were a CB on the other team's #1 or #2 WR, so to speak.

Except for the fact that he's not playing much deep safety, so it can be kind of important, depending on his role in a particular coverage and his matchup on that play.

 

[primetime]

As its been noted, the PFF that started as fan driven focus group years ago, isn't what it is now.

I think it can be taken seriously, BUT also with a grain of salt since they don't know the calls or how a specific look has been coached. AND of course, there is the matter of subjectivity that comes with giving a grade on every play.

So PFF is just a good metric for giving us an INDICATION of what a player is doing, based on the limited information at hand. It's data that we can use or ignore as we see fit. But at this point, DI, it would be wrong to say there analytics shouldn't be taken seriously. Is there another organization that does it better, or are all the analytic sites simply BS by definition.

If there are other groups that do it better, maybe we should look at THEIR grades and compare them.

In the end the only grades that count are the ones the staff gives out on a weekly basis. They are the only ones that have the requisite information, and even then subjectivity plays a role, as it was noted.

BTW- I thought they did a decent overall job. They essentially confirmed what we already knew. The Pats are made up of few stars and a lot of better than average players. We should also remember that the game of football is the one game where the sum of the individual players CAN be a lot better than the individual parts. The best talent doesn't always win.

Honestly I'd prefer if they just provided a couple lines of observations on how a given player played each game instead of a grade, but that's not magical in the way numbers are.

 

[RobertWeathers]

Yes, I understand yours and @robertweathers 's point- but as a deep safety, that twitch isn't quite so important as it would be if he were a CB on the other team's #1 or #2 WR, so to speak.

Do you see DMC as a deep safety? maybe sometimes but he seems to play a lot more of a swiss army knife role here.

 

[WaterfallJumper]

I only dislike PFF when they disagree with me.

What I find fairly useful from PFF is not their grades, but their numbers before they've assigned a grade, such as passer rating against. This article, for example, has some good information despite their usual push for grades, which I can forgive as an attempt to market their brand. They've been pretty successful at doing that, so I have to give them a hat tip from a business perspective.

The Patriots' secondary has what it takes to carry the team through the playoffs | NFL Analysis | Pro Football Focus

 

[1960Pats]

It's kinda like the Celsius-Fahrenheit thing. For some people, "average" = " bad" in the same way that 32F = "effing zero!" Celsius.

Ergo, "above average" = "mediocre", and "good" = "average".

9/5C + 32 = F
5/9 (F-32) = C

 

[1960Pats]

I have a question.

Would Steve Belichick's scouting system be useful in grading NYFL players?

 

[PatsDeb]

I just checked to confirm that Malcolm Brown wears No. 90. I am no stats guru by any stretch of the imagination, but yes, my eye test matches their stats - he is THAT BAD. You know that gaping hole that appears in the middle of the line in the run game inexplicably during most, or at least a portion of most, games this year? The hole is always right next to No. 90.

 

[luuked]

So PFF is just a good metric for giving us an INDICATION of what a player is doing, based on the limited information at hand.

The problem is that they have absolutely no way to account for anything that is even slightly out of the box / far away from the average. They are grading all players according to the strict definition of what their role on the field should be and not to how effective they are in the individual systems in which they are playing.

It is pretty much the same issue as with draft grades. They are all relative to what the "ideal" player in a certain role should look like and do. But the roles are interpreted differently by various coaches.

All that being said I think people should just ignore the ratings that PFF gives out and think more it in tiers. That seems immensely more useful to me than arguing whether a 89.8 or a 91.2 player is the better one.

 

[JMC00]

yeah I have been trying to make the point that Guy is having a damn good year (for who he is) for a while now. Only to hear that he is mostly a JAG from others.

They are the same people that think sacks is the only thing to measure DL pass rush and think Pats should let Flowers walk