Who Is the Best Cornerback: Do YPA’s Mimic the SAT?
Posted: July 14, 2008 @ 1:32 pm
The third story in BSR’s series exploring the question, “who is the best NFL cornerback,” looks at the value of K.C. Joyner’s YPA statistic. Does it measure innate cornerbacking skill, or does it measure performance?
The Educational Testing Service, the creators of the SAT exam, long maintained that their test was uncoachable. It, like all aptitude tests, claimed to measure innate ability, what an individual was capable of achieving, not what a particular student had learned. Therefore, cramming was deemed pointless, since it would not affect a student’s outcome. This belief has been exploded in recent years, as the test has been exposed as having a clear ideological bias.
K.C. Joyner has never claimed that his yards per attempt cornerback stats measure a corner’s aptitude. They are derived from pure empirical study; Joyner looks at tape of each NFL corner each year and measures the number of throws aimed his way, the results of each throw and the number of yards gained against each player. The results would seem transparent.
Nevertheless, readers of this blog reacted to my initial posting of corner YPA averages as if the stat did measure aptitude. Several readers questioned the stats, in part because the top ten CB list did not include Champ Bailey, the Denver corner considered the game’s best by many pundits.
Bailey missed the cut, though not by much (he ranked 14th among CBs with YPAs for every one of the last four seasons). He’s lower because he lacks consistency. Here’s his line, with his overall rank by year in parentheses:
| Player-team | 2004 | 2005 | 2006 | 2007 | Avg. |
|---|---|---|---|---|---|
| 14. Bailey - Denver |
8.6 (66th) | 7.1 (36th) | 4.7 (1st) | 7.8 (56th) | 7.1 |
As you can see, Bailey’s averages floated in the bottom half in ‘04 and ‘05 before he posted one of the best seasons recorded in ‘06. He dropped back into the bottom half in ‘07, allowing more than three yards more per attempt that year. And it’s not as if Bailey sat on an island most of the game and was then surprised when an opposing QB deigned to throw his way. He was targeted 60 times, only marginally better than counterpart Dre Bly, who saw 78 balls, though Bly’s YPA was an awful 8.8.
Looking at this line, I see a mediocre cornerback, with three fair seasons and one incredible one. But what of that ‘06. Was it a fluke? These number suggest not:
| Best single-season YPAs, ‘04-’07 (min. 60 att.) | |
|---|---|
| 1. Shawn Springs, Wash. -’04 | 4.2 |
| 2. Dre Bly, Detroit - ‘04 | 4.4 |
| 3. Champ Bailey, Denver - ‘06 | 4.7 |
| 3. - Fred Bennett, Houston - ‘07 |
4.7 |
| 5. Sheldon Brown, Phil. - ‘04 |
4.9 |
These are the only campaigns by every down corners the past four years to average under five yards per attempt. I think what we might be seeing are a general measure of a corner’s innate capabilities. These are the guys who get 1600s in the CB SATs. In Bailey’s case, he may not be the best cover corner year in and year out but his ‘06 shows that he’s more than able to play at an elite level.
In terms of using YPA as a heuristic, we can see that a YPA in the 4s signifies a sublime season, like a pitcher with an ERA around 2.00 or a hitter with an OPS above 1.300.
So Champ Bailey can be deemed the best corner in the game if we consider peak performance. But isn’t consistency the hallmark of a top player? Would you prefer a guy who can be incredible but plays that way only once a presidential administration, or somebody who can play very well year after year? Compare Bailey to the corners with the two best YPA averages since ‘04 and see how consistency ranks higher than flash:
| Player-team | 2004 | 2005 | 2006 | 2007 | Avg. |
|---|---|---|---|---|---|
| 1. Springs - Washington |
4.2 (1st) | 5.5 (3rd) | 6.2 (15th) | 6.2 (15th) | 5.5 |
| 2. Newman - Dallas |
5.8 (8th) | 5.8 (7th) | 7.1 (29th) | 6.1 (14th) | 6.2 |
To answer the question, I’m sure the Denver Broncos are happy to have Bailey on one edge of their defense, but if you ask Dallas fans, they’ll gladly take Terence Newman. He bounced back in ‘07, despite his plantar fascia injury and is a good bet to jump back into the top ten with more help in the Dallas secondary.
Another aspect of YPAs that deserves mention here is how they remake the term “shut down corner.” Fans take this term literally; when a blogger posted Dominique Rogers-Cromartie’s college stats, which showed opponnents completing less than 30% of passes in his direction, another blogger belittled those numbers, saying they didn’t reveal any special performance.
In fact, if an NFL corner posted Rogers-Cromartie’s line on a consistent basis he would get fast-tracked to Canton.
Stopping opponents 50% of the time appears to be the threshold for being considered a shutdown corner. In the last three years only 38 corners have achieved this — 9 in 2005, 17 in 2006 and 12 last season. Since shutdown connotes performance far above 50%, I think that the term be discarded or ignored; shut down corners simply don’t exist.
Only once in that span has a corner topped 60% in success percentage, that coming last year when the Raiders Nnamdi Asomugha posed a 62.9% success rate. Asomugha is the hardest corner to throw against, topping 50% in each of the last three years, the only cornerback to do so.
Might Asomugha then claim to be the best corner? Perhaps, but his YPAs indicate that while it’s hard to complete a throw against him, teams gain large chunks of yards when they do.
The YPA is hardly a perfect cornerback stat, but playing with YPA stats reveals a new cornerback world, one that elevates the steady over the flashy. Corners with 1600 SATs are impressive, but it’s the guys with scores in the 1200 and 1300s, the cornerback nerds who grind day-after-day and year-after-year who have the most success.
Comments
12 Responses to “Who Is the Best Cornerback: Do YPA’s Mimic the SAT?”
Got something to say?
You must be logged in to post a comment.
-
"The Sports Doctors"
iTunes/
Winamp
Windows
Media
RealOne -
Cowboy Up
Dedicated to providing the best Cowboys analysis anywhere, our efforts are free of charge but not free of expense. Your kind donation today will keep our drive alive.
Thank you
Or mail payment to:
RV2 Enterprises
P.O. Box 150093
Austin TX 78715-0093
Advertisers
OnlineSeats.com
Dallas Cowboys Tickets
Dallas Mavericks Tickets-
Fave it!
-
SiteMeter
Burst Media
The new SAT has 2400 points, by the way.
and how many people have taken it?
You always get the last laugh, Raf.
The Ranch almost never gets the best of the Vela in terms of analysis, and the owners of that blog didn’t in this case either. But a poster there — ADAM JJT — put up an analysis that I think beats the excessive focus on YPA as the measure of a CB.
I linked to it in the last post, but it’s worth re-posting here.
http://mbd.scout.com/mb.aspx?s=112&f=1786&t=2671390
Plus, this thread has the added value of giving us some information about Adam Jones, which Vela’s three posts on this subject have missed entirely.
So, Rafael, I generally like your work, but I think you’ve given us a flawed perspective on CBs in this 3-part series. Shawn Springs, for example, comes out smelling like roses in your series, but drops down to a more realistic rating when other factors are considered.
The good news is that both T New and A Jones are outstanding CBs — near the top of the league. Henry isn’t nearly so good, but he’s certainly capable enough to get on the field when 3 or more CBs are in the lineup.
VAfan,
1. who said my series is over, or that it will ignore Adam Jones?
2. I’ve contended from the beginning that YPAs are a starting point for debate, not the end point. The forums are an opportunity to discuss their limits and ways to use them.
3. I read the post. It’s interesting but I think you neglect a very important point in your zeal to puff up AdamJT13’s work — he’s basing his numbers on Football Outsiders and Joyner’s metrics. In other words, he’s not doing any film work that I can see, but reinterpreting theirs.
Now, he may well have a better mousetrap. He makes some very convincing points. But if you dismiss Joyner’s work because you think it’s flawed, then what does Adam have to work from?
That’s called throwing the baby out with the bathwater.
4. His study only looks at one year. I’ve taken the numbers over four and then three.
Apples and oranges.
Sorry, mister but what am I supposed to do, link to the other story and stop writing because you think it’s better?
Let Adam JT13 write his own blog. I think he did at one time.
My 2008 season tickets arrived today!
my daughter is taking the sat this year. 2400? damn. she goes to a school with some BS 7 point grading system too (its a good school but their grading pisses me off).
Someone mentioned getting press credentials for rafael recently. Let me state for the record that I am all for it. this is one thing i would love to see more bloggers get, especially those with quality sites like this one.
Raf,
Has KC ever run metrics on Deion Sanders? I’m wondering now how much of a “shut-down” corner he was?
DaveMorgan,
I can hear a coach talking to his QB now, “c’mon, just throw the ball prime’s way. his metrics say he’s not a shutdown corner, we’ve got nothing to worry about.”
Raf,
You don’t have to get testy. I like your work. You beat the other guys almost all the time. I just wouldn’t rate this series as among your best. But it’s a tough nut to crack, too, so at least I admire your attempt.
As for Adam Jones, I asked several times if you were including anything and didn’t get a peep from you. This is your first hint that it’s coming unless I missed something. I’ll be interested if and when you finally include it.
VAfan
Yeah, that’s kind of what I was wondering, jarhead. I don’t know what his metrics were, but Deion is someone I’d never throw on very much.
Hey, the reason i crushed the SAT’s was due to its ideology, at least according to that article at Gladwell.com. I knew as a kid that multiple choice was a silly way to truly test, and provided me with a sweet advantage… for instance, in the math portion, you can eliminate the outlying answers and work backwards from the likely answers.