Post by dachuckster on Oct 7, 2023 22:36:20 GMT -6
What teams get out of PFF, has nothing to do w/what random people outside of the org can do with it. Again we, nor PFF, know WHAT they are asked to do, and they do not just grade on the obvious; if they did, they'd be no better then random drunk fans watching on TV.. They've stated as much. Keep trying to defend that terrible org though, it's glory days are well past and they haven't even been around that long.
The plays you are talking about where a receiver runs the wrong route, a lineman forgets his assignment, etc. are only a subset of plays where you could give a wrong grade because you don’t know the “ask.” I believe the PFF people have addressed the issue, and say they don’t grade the player on a play if there is clearly some ambiguity about his assignment on the play. On most plays, it is not hard to see what he is trying to do and hard to imagine him “asked” to do something different. It varies by position.
As a former professional sociologist trained in multivariate statistics, I would like to know how much “predictive validity” a player’s PFF grade has. For example, how well is it statistically correlated with other events, such as:
>Moving up/down on the depth chart.
>Getting cut before next season
> Salary and contract
>Getting voted to Pro Bowl
>Getting a contract extension after rookie contract expires (I especially like this one.)
If there is a statistical correlation at the .01 confidence level, I would conclude that PFF score correlates pretty well with playing ability – not necessarily for a single game, but over a season.
I would like to add if anyone ever did any real analysis on seeing if all of their analysts would rate similar plays in similar situations the same way. I personally suspect that the process by an analyst has some element of subjective judgment which might make it difficult to standardize their process/product.
