This article by Barry Petchesky is a lazy mash-up of two headlines that are lazy in and of themselves: "Rex Ryan Don't Know Much O-Fense" and "Quarterback Rating is an Overrated Statistic" are hackneyed on their own; cobble them together and what do you get? A messy, masturbatory exercise that essentially comes to the conclusion that statistics are useless because Rex Ryan likes them, which: huh?
Petchesky basically takes the position that Ryan... wait. Let's take a look at the position Petchesky is taking:
Yikes. OK first things first. [hops on internets for 31 seconds].
Adum Dunn - Career OPS .861 / 2013 OPS .762
Brett Gardner - Career OPS .733 / 2013 OPS .759
So we can safely assume Petchesky means 2013 OPS. We should point out that Dunn's season was one of (likely) decline, while Gardner in turn slightly overperformed his career marks. Much more importantly: this analogy illustrates the point (completely undermining what we can only conjecture is Petchesky's actual point) that statistics which aggregate other statistics, while quick, dirty, and imperfect, can help us draw meaningful comparisons between players or teams that may otherwise seem apples-to-oranges. No one who has ever watched baseball would ever argue that Gardner and Dunn are "similar players," but (gasp!) based on their OPS numbers this year, one might say that Gardner and Dunn were more comparably productive, offensively, than they had been in previous years. That's the beauty of aggregating statistics! And Rex Ryan loves them! That, of course, is what we're talking about with this Adam Dunn/Brett Gardner analogy, right? Right??
To go one step further toward doing Petchesky's job for him:
Brett Gardner OBP .352 / SLG .381
Adam Dunn OBP .320 / SLG .442
Gardner and Dunn both put up very similar OPS numbers in 2013, in very different ways: Dunn slugged his way to his ~.760, while Gardner OB'd his way to his. Again, this deconstruction of these statistics could/would have been the "meat" of his article, but he glossed over the part that was the, er, "point."
To go two steps further toward doing Petchesky's job for him (is someone paying me yet?) (**this time using ultra-novel ACTUAL NFL FOOTBALL QUARTERBACK STATISTICS!):
Player: TD/INT/YdsPerAtt/Comp%/Rating
7. Locker 6 / 0 / 6.50 / 62.2 / 99.0
8. Wilson 11 / 4 / 7.96 / 61.5 / 97.2
9. Cutler 12 / 6 / 65.9 / 7.51 / 95.2
From this breakdown of traditional Quarterback Rating, we can say things like "Quarterback Rating overvalues INT's, as Locker's zero picks puts him in the top 7 despite a weak 6.5 yd/att," or "Quarterback Rating might lead you to believe Wilson and Cutler are totally comparable players, but this ignores Wilson's rushing numbers." Again, Petchesky does not do this here. He talks about how statistics only tell us what happened after the fact. In baseball.
It's tough to find a stat that's predictive. Giving significant weight to passer rating is akin to saying a quarterback is good because his quarterbacking stats are good. It isn't much more than a tautology.
Yes, dingbat. Statistics allow us to keep track of and evaluate what happened in the sports that we watch. That is their ontological purpose, vis a vis sports. We can then decide which players to play, based on their performance, on the assumption that they can continue their performance. Why am I saying this? Oh, cause we're saying stupid, meaningless stuff? OK.
To kick a dead horse while he's down (or something): Petchesky points out that quarterback rating "elides contributions from the other 21 players on the field." But to hear Rex tell it, as he did in the post-game, you'd notice he has an almost holistic approach to QB rating. He seems to understand that if his team has a good game, then his QB will have had a good rating. Petchesky does not seem to understand that Rex understands this. Or (more likely) he refuses to acknowledge it and writes what he wanted to write anyway, because it's easier and fits the Rex Ryan narrative. How very ESPN-y of you, Barry.
The point is that if the NYT source article had been about, say, "number of plays" and "Chip Kelly," Petchesky would not be writing this article denigrating statistics as "unable to predict the future," but since Rex was finally agreeing with the VORPies (and again, not really - it's not like Rex was talking Football Outsiders' DYAR or even ESPN QBR ), Petchesky found it the right time to blast them. I hate to defend Rex and his recent discovery of statistics, but this, I think, is even worse. What a big mess.
Petchesky basically takes the position that Ryan... wait. Let's take a look at the position Petchesky is taking:
"The amazing thing is at first, I was like, What is this?” Ryan said,
referring to his understanding of passer rating. “It wasn’t that big a
deal to me."
That was until Ryan realized the correlation between a higher passer rating and victory—fairly steady around 80 percent over recent seasons.
That was until Ryan realized the correlation between a higher passer rating and victory—fairly steady around 80 percent over recent seasons.
The
correlation isn't a surprise. The passer rating formula takes into
account completion percentage, yards per attempt, TD percentage, and INT
percentage. If you win those battles, you're generally going to win the
war.
I don't understand. We're holding it against Rex that he values a statistic that measures 'those battles that you need to win to win the war'? If anything, we should be lauding Ryan for finally taking an "analytics-based" (lol) interest in the offensive side of the ball.
But while compressing four separate statistics into a single one
makes for simple shorthand, it tells you less about a quarterback than
if you consider each one individually.
If Petchesky were going to write an interesting article, it would have happened here. How, precisely, does considering yards per attempt and TD percentage separately give the fan a clearer understanding of how his/her quarterback, or team, played in a given sample? The answer to that question would make for a (mildly) interesting article. That article is not what Petchesky wrote.
But here's the kicker:
But here's the kicker:
It's like assuming Adam Dunn and
Brett Gardner are similar players because they have near identical OPS
figures.
Yikes. OK first things first. [hops on internets for 31 seconds].
Adum Dunn - Career OPS .861 / 2013 OPS .762
Brett Gardner - Career OPS .733 / 2013 OPS .759
So we can safely assume Petchesky means 2013 OPS. We should point out that Dunn's season was one of (likely) decline, while Gardner in turn slightly overperformed his career marks. Much more importantly: this analogy illustrates the point (completely undermining what we can only conjecture is Petchesky's actual point) that statistics which aggregate other statistics, while quick, dirty, and imperfect, can help us draw meaningful comparisons between players or teams that may otherwise seem apples-to-oranges. No one who has ever watched baseball would ever argue that Gardner and Dunn are "similar players," but (gasp!) based on their OPS numbers this year, one might say that Gardner and Dunn were more comparably productive, offensively, than they had been in previous years. That's the beauty of aggregating statistics! And Rex Ryan loves them! That, of course, is what we're talking about with this Adam Dunn/Brett Gardner analogy, right? Right??
To go one step further toward doing Petchesky's job for him:
Brett Gardner OBP .352 / SLG .381
Adam Dunn OBP .320 / SLG .442
Gardner and Dunn both put up very similar OPS numbers in 2013, in very different ways: Dunn slugged his way to his ~.760, while Gardner OB'd his way to his. Again, this deconstruction of these statistics could/would have been the "meat" of his article, but he glossed over the part that was the, er, "point."
To go two steps further toward doing Petchesky's job for him (is someone paying me yet?) (**this time using ultra-novel ACTUAL NFL FOOTBALL QUARTERBACK STATISTICS!):
Player: TD/INT/YdsPerAtt/Comp%/Rating
7. Locker 6 / 0 / 6.50 / 62.2 / 99.0
8. Wilson 11 / 4 / 7.96 / 61.5 / 97.2
9. Cutler 12 / 6 / 65.9 / 7.51 / 95.2
From this breakdown of traditional Quarterback Rating, we can say things like "Quarterback Rating overvalues INT's, as Locker's zero picks puts him in the top 7 despite a weak 6.5 yd/att," or "Quarterback Rating might lead you to believe Wilson and Cutler are totally comparable players, but this ignores Wilson's rushing numbers." Again, Petchesky does not do this here. He talks about how statistics only tell us what happened after the fact. In baseball.
It's tough to find a stat that's predictive. Giving significant weight to passer rating is akin to saying a quarterback is good because his quarterbacking stats are good. It isn't much more than a tautology.
Yes, dingbat. Statistics allow us to keep track of and evaluate what happened in the sports that we watch. That is their ontological purpose, vis a vis sports. We can then decide which players to play, based on their performance, on the assumption that they can continue their performance. Why am I saying this? Oh, cause we're saying stupid, meaningless stuff? OK.
To kick a dead horse while he's down (or something): Petchesky points out that quarterback rating "elides contributions from the other 21 players on the field." But to hear Rex tell it, as he did in the post-game, you'd notice he has an almost holistic approach to QB rating. He seems to understand that if his team has a good game, then his QB will have had a good rating. Petchesky does not seem to understand that Rex understands this. Or (more likely) he refuses to acknowledge it and writes what he wanted to write anyway, because it's easier and fits the Rex Ryan narrative. How very ESPN-y of you, Barry.
The point is that if the NYT source article had been about, say, "number of plays" and "Chip Kelly," Petchesky would not be writing this article denigrating statistics as "unable to predict the future," but since Rex was finally agreeing with the VORPies (and again, not really - it's not like Rex was talking Football Outsiders' DYAR or even ESPN QBR ), Petchesky found it the right time to blast them. I hate to defend Rex and his recent discovery of statistics, but this, I think, is even worse. What a big mess.
No comments:
Post a Comment