Beware PFF grading system

[Erebus]

I have a subscription to Pro Football Focus, and I've read their grading system explanation, but it just doesn't add up to me, and they don't respond to my emails when I call them out on their BS.

Based on their grades, Richard Sherman currently is the 8th ranked CB in the league in terms of coverage, while allowing a passer rating of 22.9. Ahead of him are Janoris Jenkins (124.2 passer rating) and Jason McCourty (106.2). The rest of their stats fall in line with the passer ratings they allow.

It appears PFF is completely subjective when giving grades on each play. Players don't get any points for covering someone so well that the QB won't even consider throwing at you.

As many of you know, Earl Thomas was graded as the 34th best safety in the league last year. It all seems bogus to me, and I will not be renewing my subscription this year.

Maybe someone with a better understanding of PFF can clear this up for me.

 

[MontanaHawk05]

I've never bought their ratings system. They've even acknowledged the numerous accusations from people that they're just making up a contrarian system (and heavily advertising it to the point of annoyance) in order to get site hits.

 

[RolandDeschain]

I'm a premium subscriber at both PFF and Football Outsiders. I would not say PFF's CB rankings are as bad as you say. Their safety ones are, (particularly toward Earl) but the CB ones are not as bad.

The main reason this looks so jacked up is because it's still early in the season, Erebus. PFF does not account for opponent difficulty the way Football Outsiders does. Corners are at a disadvantage in PFF ratings when facing good receivers. It will even out over time. Check back in a few weeks. Right now, half the names in the top 5 CBs on PFF are there because of having faced scrubs thus far.

Sherman was #2 in the NFL last year per PFF, only Antoine Winfield narrowly edged him out; and most of Winfield's score came from defending the run, not coverage.

Again, though, they do need to improve their formulas. I'm not defending them, just explaining more.

 

[Smelly McUgly]

I have a hard time with PFF, harder than I do with some of FO's stuff. At least FO heavily leans on DVOA for evaluating teams, which is something that I can both understand and trust.

 

[RolandDeschain]

DVOA is hands down the best in the NFL business for stats. No comparison.

I like PFF's stats for the ease with which I can find certain types of information, and one thing to keep in mind is that all teams benefit or are penalized by the same problems in PFF's formula; so while you may see a rather inaccurate rating, you can still use where it is compared to others on the list in a somewhat reliable way MOST of the time.

That being said, you can never trust stats in the first few weeks of each season. Too many anomalies that even out over time.

 

[kearly]

I think the problem is that you have different people scoring different games. Imagine if in the Olympics they only had 1 judge give the score and that judge changed for every contestant. Wouldn't exactly be very fair or precise, would it?

I'm sure some people doing the scoring know exactly what they are doing, while others are clueless. Evaluation is a talent. Some have it, most don't.

 

[JSeahawks]

This is why I don't pay any attention to all these stat sites. I just go off the eyeball test.

And by that I mean Kearly's eyeballs.

 

[RolandDeschain]

That's a very valid point, Kearly. However, Football Outsiders does the same thing in that regard; many different eyes. They certainly do it a lot better, though.