Football outsiders: best CBs in 2014 (with Gaines at 9)

Best Cornerback Stats 2014

by Andrew Healy

http://www.footballoutsiders.com/stat-analysis/2015/best-cornerback-stats-2014

For this year’s edition of the cornerback charting stats, we are conducting the festivities Academy Awards-style. We will be handing out Best Cornerback awards both in a leading role (No. 1 corner) and a supporting role (everyone else). As with the Oscars, these distinctions are imperfect. The leading corner is not clearly defined on every team. But with all the work we do to adjust for the quality of opposing receivers, it ends up making more sense to try to separate the Darrelle Revises and Richard Shermans of the NFL from the Orlando Scandricks and Corey Grahams (much more on him to come).

As in the past, the cornerback charting stats come from the Football Outsiders Game Charting Project. We track two main stats in particular, adjusting each for the quality of opposing receivers:

Yards Per Pass, which is just the average yards the corner allowed when targeted.
Success Rate, which is the share of targets on which the corner prevented a successful gain (45 percent of needed yards on first down, 60 percent on second down, and 100 percent on third down).

We are identifying the plays in which the cornerback is listed as the primary defender in pass coverage, ignoring a few types of plays unlikely to reflect on the cornerback. Specifically, we ignore screens, balls tipped at the line or thrown away, Hail Marys, and plays where the quarterback was hit while throwing. We include pass interference, but ignore other defensive penalties that often occur away from the pass.

We also report the number of targets that each cornerback faced, as well as a metric we call “estimated target percentage.” This stat uses snap counts to estimate what percentage of possible targets were thrown at this player when he was on the field. The “possible targets” part of that metric leaves out the passes noted in the previous paragraph, as well as those passes listed with “Uncovered” or “Blown Coverage,” though the total of possible targets does include “Hole in Zone” passes.

.

The usual caveats about game charting data apply here, as described in last year’s article. As always, we need to point out that this is imperfect data charted by a group of volunteers plus a handful of FO staffers. Cornerback stats have been very volatile, bouncing up and down year-by-year for a lot of players, and the best cornerbacks will sometimes rank lower than expected because quarterbacks only throw in their direction when they make a mistake. In addition, with so many players ranked, readers need to make sure to understand that there isn’t much difference between ranking 35th and 40th in a metric. For that matter, there isn’t much difference between ranking 35th and 50th. As we always say, these stats should not be seen as absolute statements on player value. They’re just part of the story.

The stats below are adjusted for the quality of the opponent receivers. These stats compare No. 1 receivers to other No. 1 receivers, No. 2s to other No. 2s, and so on. We are considering changing these adjustments in the future, perhaps by comparing receivers being covered to all receivers rather than just other receivers of the same “receiver position,” but we couldn’t do that for 2014 due to time constraints. Nevertheless, the current adjustments provide interesting insights, particularly when we break down the data by separating No. 1 corners from other corners.

As described earlier, the No. 1 corner is more obvious in some cases than in others. Some teams use their starting corners interchangeably, while other teams saw playing time scrambled by injuries. For this article, we’re defining a No. 1 corner as the one on each team who faces the most opposing targets against opponents’ No. 1 receivers. This definition usually gives the most logical No. 1 corner, but not always. Even when a team has a clear top corner, he sometimes does not always cover the top receiver (e.g., Kyle Arrington rather than Darrelle Revis guarded T.Y. Hilton for most of the AFC Championship Game, with help from Devin McCourty on almost every play). Still, over the course of the season, the list of top corners mostly accords with common sense. We made two changes (naming Vontae Davis as the Colts’ top corner over Greg Toler, and Chris Harris for the Broncos over Aqib Talib), but left the rest defined according to targets. This definition makes Lardarius Webb the Ravens’ top corner, not Jimmy Smith, whose season was cut short by a Week 8 injury.

Note that to qualify in either category a corner had to either face 50 targets or start eight games.

Best Cornerbacks in a Leading Role

The chart below has the ten top contenders for the best No. 1 corner in the NFL in 2014. We also included two players who would have made the top ten if we had restricted the sample to players who made at least 14 starts. Ranks represent placement among the 32 cornerbacks we considered as No. 1 corners.

It is a pretty close race at the top, but Darrelle Revis wins top honors. On top of having the highest adjusted success rate, Revis was also the least targeted of the No. 1 corners. Revis was targeted on just 17 percent of opponents’ pass plays, and he gave up a successful play on 38 percent of his targets. Multiplying those together, we see that Revis gave up a successful play on only 6.5 percent of opponents’ throws when he was a possible target. This number may do a pretty good job of identifying the consensus top corners in football. Chris Harris, who ranks third by adjusted success rate, comes second to Revis by that metric, surrendering a successful play on just 7.1 percent of opponents’ pass plays. However, Harris’ numbers may overrate him at least a little bit given his responsibilities in the Broncos’ scheme compared to the responsibilities of Revis and Richard Sherman. Sherman, fifth by adjusted success rate, comes just behind Harris at 7.2 percent.

In addition to delivering some memorable hits, Vontae Davis also took his coverage to another level in 2014. While teams did not avoid Davis the way they avoided someone like Revis, Davis came close to matching Revis by adjusted success rate. Still, Davis only faced the player we identified as his opponents’ top wideout on 41 percent of his targets, compared to 61 percent for Greg Toler. (As he did with Chris Harris in the Film Room column linked in the previous paragraph, Cian Fahey also analyzed this issue in Vontae Davis’ game in an October article.)

Coming in sixth is Brandon Flowers, the cornerback who the Chiefs did not want. Regardless of his size, Flowers was a bright spot in a poor San Diego defense. The Chargers ranked 25th overall in defense, but 13th against opponents’ No. 1 wideouts. The nominee with the most potential to move up these charts might be the Vikings’ Xavier Rhodes. In just his second season, Rhodes ranked seventh in adjusted success rate. Against opponents’ top wideouts overall, the Vikings were just 26th by DVOA, but that reflected their hopelessness when anyone other than Rhodes was in coverage. While Rhodes had an adjusted success rate of 54 percent in his 45 targets against opponents’ top wideouts, the Vikings’ other corners had a disastrously low 35 percent adjusted success rate in that role, a number that is lower than the overall adjusted success rate for the worst number one cornerback in football.

Who is that player? We’ll tell everyone in another article next week that will look at the worst cornerback stats of 2014. If you read our Four Downs piece on weaknesses in the AFC East, you won’t be surprised when it turns out not to be Bradley Fletcher.

Best Cornerback in a Supporting Role

The Best Supporting Cornerback for 2014 was a slam dunk. Corey Graham, part of a Bills defense that ranked second in DVOA, blew away the field. Ranks here represent placement among 45 cornerbacks who didn’t qualify as No. 1 corners but had either eight starts or 50 charted passes.

Graham posted a 71 percent success rate, nine points higher than any other corner in football, and allowed just 4.8 adjusted yards per pass, almost a full yard better than any other corner. With Stephon Gilmore one spot off the list for No. 1 corners, the Bills’ secondary is perhaps underrated. They benefit from a pass rush that ranked first in the NFL in adjusted sack rate, but the secondary leaped off the film on its own — against Green Bay, to give one strong example. Graham’s dominance in his role extended far beyond that game.

Other standouts include Orlando Scandrick, who played well enough to deserve his recent contract extension. Brandon Browner ranked seventh in adjusted success rate. The Patriots’ tandem of Revis and Browner edges out Denver’s Harris and Aqib Talib (ninth for supporting role corners in adjusted success rate) as the best pair of cornerbacks in the NFL last year. Now both are gone to other teams.

I extended this list to the top 15 corners to note the interesting replacement going on in Seattle. The Seahawks swapped out the corner ranked 15th in adjusted success rate (Byron Maxwell) for the one ranked 13th (Cary Williams). With Maxwell receiving 3.5 times more guaranteed money, the Williams contract looks like the much smarter deal. Williams allowed a much higher adjusted yards per pass, but he would likely look better on that account if he had been the one with Earl Thomas playing in support. On the other hand, these Williams stats demonstrate some of the inconsistency we’ve found with cornerback charting metrics. Until 2014, Williams had always ranked much better in yards per pass than in success rate. In FOA 2014, we joked that “Cary Williams has a large tract of land in front of him, and he would like to invite you to visit.” Last year, his stats showed the exact opposite to be true, as Williams was more consistently successful but burned by deep throws.

All the cornerback charting stats for 2014 can now be found on player pages. Note the rankings will differ from those above because the rankings on the player pages do not separate out No. 1 cornerbacks. Also, the target percentages will be different because they are based on games played, not snaps, to match target rates for charting data prior to 2013. At some point after FOA 2015 is finished, we’ll be fixing those player page numbers to match this new way we are determining target percentage by considering snap counts.