Bennett. Fred Bennett. Revisited.
Jul 21, 2008 Demarcus Faggins sucks, Dunta Robinson, Faggination, Secondary issues are primary, The Fred, The Future, Training camp 2008
In response to the article quoted in this post, I emailed KC Joyner and asked for a little more explanation of the methodology. He was kind enough to reply.
Try this on for an expansion:
In Scientific Football 2008 and my 2008 Draft Guide (both of which can be ordered at www.TheFootballScientist.com) I rated matchups by color ranking, with red being difficult, yellow being average and green being favorable. Against red matchup WRs in 2007 (those that gained 9.0 YPA or higher), Bennett allowed only 6.8 YPA. That’s very good (29th best out of 95 qualifiers) but it pales in comparison to how Bennett handled yellow rated (7-9 YPA) and green rated (less than 7 YPA) WRs. His 3.3 versus yellow rated was the 2nd best in the league and his 3.0 against green rated was tied for 19th. Add them all up and his overall YPA against WRs was 2nd best in the league. I’ve been touting him as a future Pro Bowl candidate for how well he played last year.
So, hopefully, that clears it up a bit more. Bennett was good against top-notch receivers, very good against mediocre guys, and downright dominant against the guys in-between. The good news being, of course, that if you assume talent is distributed on a bell curve, Bennett (and all corners) are going to face a lot more “yellow” receivers than “red” or “green” ones.
Now, I know this seems foreign to us as Texans fans, but IF Molden really is ahead of where Bennett was this time last year (as sources seem to be saying) AND IF Ray Rhodes really saw enough talent in Jacques Reeves that the organization thought Reeves was worth that price tag AND IF Dunta comes back at even 90% of where he was…well, kids, we might just be in the enviable situation of having too many very good (or better) CBs. (And, as a bonus, we can finally give Petey his walking papers.) Even better, with only one or two of those IFs coming to fruition, we are finally set at CB.
This is all foreign and slightly frightening. Hold me.
(click arrow to reply)
July 21st, 2008 at 9:53 am
Seems awfully qualitative to me, the grading of WRs.
But hey, I hope he’s right.
(click arrow to reply)
July 21st, 2008 at 9:54 am
I know I’ll probably get a lifetime ban for this, but….
….
….
….Petey is not a bad nickle back. He was horribly mis-cast as a starter, obviously.
*ducks*
(click arrow to reply)
July 21st, 2008 at 9:59 am
@bigfatdrunk: I’ve decided that (a) this may not have ever been true, as it’s not like we were running a great secondary out there when Faggins was the nickel, so he might just have not been as noticeably bad when on the WR3, (b) it certainly is true that he never should have been a starter, and (c ) he showed last year when moved back to nickel that he sucks horribly at that position as well now.
(click arrow to reply)
July 21st, 2008 at 10:02 am
@TheBlowLeprechaun: Yeah, I’m not sure where the <7, 7-9, and >9 categories came from unless that is some sort of bell-curve distribution as well. Still, I’m with you. I hope the dude is right on. (I also think it’s awesome that Fred completely shutdown that middle group, however you want to define them.
(click arrow to reply)
July 21st, 2008 at 10:07 am
@Matt: Sigh. True enough, though I do wonder how beaten down his confidence was after being so thoroughly toastable last year. Still, as long as he’s not a CB1 or CB2, I’m a happy camper.
(click arrow to reply)
July 21st, 2008 at 10:34 am
Two thoughts;
1. I hope this post is not jinxtastic. I think it has sufficient qualifiers. For all of our sakes, I hope this isn’t one of those blog posts that you look back at and start laughing out of pain because the secondary is still toasted. I’ve written a few regrettables.
2. I love Pete. Just not as a starter. I’m never one to bash try hard role players who were put in positions to start because the front office never got someone better. (Bennett was hurt early in the season, and Dunta liked playing with Pete. Maybe perhaps because it made his job easier because all the balls would go Pete’s way. I think J.P. Losman is still throwing Pete’s way).
(click arrow to reply)
July 21st, 2008 at 10:36 am
If I may be slightly nerdy for the moment, not having seen the actual data, he most likely did something as simple as looking at all receivers and then calculate the top 10% or so and figured where the cutoff point was (then rounding to a nice neat number like 9). He then likely did the same thing for the “green” receivers.
-
That’s not particularly exciting, but here’s what IS exciting. It’s actually pretty unlikely that receivers are normally distributed (geek term for bell curve), but more likely that there is some positive skewness. All that really means is that there are likely more “green” receivers than “red” receivers. Still, though, the overwhelming majority of receivers are likely to fall in the “yellow” category. Probably somewhere in the range of 70-80%. Add in the “green” receivers and The Fred was absolutely dominant against roughly 90% of the league and well above average against the remaining 10%.
-
The other thing he might have done is divide all receivers in to equal thirds and then base his cutoff points on the closest “neat” number to each group. Even still, this would suggest that The Fred was dominant against 67% of the league and above average against 33%.
-
This approach seems less likely to me though, but what makes all of this even MORE exciting is that Bennett was only a rookie last year and now has more experience AND Ray Rhodes. Consider me excited.
-
/dork
(click arrow to reply)
July 21st, 2008 at 11:50 am
I think the statistic becomes more meaningful when coupled with quarterback sacks/hurries. My recolletion was that Dunta’s playing term was before Mario shifted gears and Fast Freddie’s was after. Does this skew the statistical equation?
(click arrow to reply)
July 21st, 2008 at 12:24 pm
@da Bull: That’s a fantastic point. I will follow up with Joyner once more and see if I can get an answer. Also, assuming you are correct, that means that, as our pass rush improves, Bennett will be even BETTER. Rock.
-
@Vega: Nerd. (Thanks for clearing that up.)
(click arrow to reply)
July 21st, 2008 at 1:01 pm
@Matt: Yes to the point of better pass rush = better Bennett. But when making the statement of Bennett being statistically better than Robison what variables do you include? Pass rush? Who is the opposite cornerback and the associated safety? My point being, to pull a number out of your butt and say that your point is proven is a bit misleading….although I like what he is saying so I’ll put my blinders on to the other variables.
(click arrow to reply)
July 21st, 2008 at 8:54 pm
The Fred passes the eyeball test. If Molden has half the impact Fred had, we’ll be fine (assuming Reeves doesn’t suck).
(click arrow to reply)
July 21st, 2008 at 9:33 pm
@Charlie: And so long as Reeves has a pulse, he’ll be better than Sir Suck.
(click arrow to reply)
July 22nd, 2008 at 1:41 am
…and now, to nerd things up even more!
-
“Still, though, the overwhelming majority of receivers are likely to fall in the “yellow” category. Probably somewhere in the range of 70-80%.”
-
Isn’t a standard deviation in a normal distribution something like 68%? And two standard deviations approaches 95%? If that’s how the groups were color-coded, you really shouldn’t even care about the “suck” receivers so much because, well, (A) they suck and (B) there’s so few of them. Anyway judging by the numbers used, it doesn’t seem to me that he used a normal distribution, but rather just cut the groups up in to thirds. I would imagine a normal distribution would be the smarter way to go about this, but he’s the football scientist, not me.