Can Anyone Cover Anyone?

On Monday, Rich Hofmann argued in his column that the real problem with the Eagles' defense wasn't ineffective blitzing or the the four-man pass rush, but rather the guys behind the guys:

Overall, here is the key: In a similar number of snaps with a four-man rush, the Eagles had a couple of more sacks in 2009 than they did in 2008.

If that is the case, then, what accounts for the overall drop in defensive numbers and the increase in the number of big plays? It was the coverage. It is no longer good enough to support such a gambling, aggressive defensive system ...

If the Eagles were to make their centerpiece offseason acquisition a defensive end, I wouldn't boo. They need difference-makers on that side of the ball, wherever they play, period. But I would start on the back end and work toward the line of scrimmage. I'd start with a cornerback.

I thought it might be interesting to use that new statistics site I talked about yesterday to see if there was any way to validate Hofmann's argument.  A few hours later with some excel slicing and dicing and here we are.

---

As a quick note before we start, there's a chicken and egg component to the case he's making.  The term "coverage sack" was invented for a reason.  And there's no question it's easy covering guys when the front four (through seven) are harrassing the heck out of the quarterback.  But if you run some quick and dirty correlations between the various coverage (INTs, YPA, etc.) and pressure (sacks, hurries, etc.) metrics, you find that the effects mostly go the right way, but they're also not that large.  Even a great pass rush can't always cover for a terrible secondary.

---

Hofmann begins by arguing that the pass rush from the front four was pretty good this year.  That's certainly true on the outside, where Cole was great and Parker was again pretty decent.  It's not, however, really true in the middle of the line, where Bunk and Patterson are a very good run-stopping duo but aren't really allowed to get after the quarterback (click here, sort by "rush," scroll down ... keep scrolling).

Overall, the PFF graders aren't really that impressed with the Eagles' pass rush.  To get a better sense of why, compare their player rankings for the Eagles and Vikings.  The Eagles basically have one plus pass rush guy -- the Vikings have three.

In fact, there's part of me that thinks maybe there should be a bit of a role reversal on the line next year.  Assuming Antonio Dixon shows up in shape this summer and looks like he's ready to build on his strong start, why not let him be your first and second down run stopper and allow Bunk -- who really is an incredible athlete -- a little more freedom to create havoc?  Bunk could be the one run and pass guy we haven't had inside for a long time.

Still, I think the underlying point is fair.  There's plenty of room for improvement, but the rush was a lot better at generating sacks than some of the truly down years they've had in the past.

The question then becomes, if the coverage was bad, who's to blame?  And is Hofmann right that the number one priority should be cornerback?

This is when things get interesting.  Let's start with the CBs.

Here's the table of all cornerbacks who played at least 25 percent of their team's snaps this year.  The column at the far right is opponent's passer rating on throws at that cornerback.  It's not a perfect metric by any means, but it's a useful comparison.

If you click on that column header to sort by it, you'll see these numbers:

Rank - Player - Pass rating
#21 - Sheldon Brown - 68.2
#45 - Asante Samuel - 80.0
#50 - Joselio Hanson - 81.8

Those don't look so bad.  Although, one of the problems with using passer rating is that it places a high importance on interceptions.  And that makes an impact here, because no cornerback group in the league intercepted a higher percentage of passes intended than those three guys did (you won't find that on the site, that's in my Excel file).

So what happens if you do the same ranking, but you take out all interceptions, to get a better sense of what their coverages were like on plays where they weren't picking balls off:

Rank - Player - Pass rating
#46 - Sheldon Brown - 92.4
#58 - Joselio Hanson - 97.2
#96 - Asante Samuel - 119.6

There are 107 cornerbacks on the list.  Basically, if you take away the interceptions (for every CB in the league, not just these three), you get two average players and a guy who's just barely out of the bottom 10.  A similar, albeit less drastic, effect is found if you go back to that link above and rank by yards-per-reception (a variation on the infamous Lito Sheppard statistic).  Asante and Sheldon are #74 and #78 on that one.  (Lito is #82.)

Of course, you can't just pretend the interceptions didn't happen.  The average turnover is worth in the neighborhood of four points, which means you'll put up with some risk-taking to force them.  But this seems a tad extreme.

On the other hand, I'm not sure this isn't primarily a coaching issue, at least for Asante.  If he gambles too much or refuses to press the way he's supposed to, stick his butt on the bench.  That's what it's for.

Moving to the next position, the safety question is a lot more straightforward.  Basically, Quintin Mikell is good and the rest of the guys stink.

Mikell's opposing passer rating was 76.9.  Sean Jones was 109.3 and Macho Harris was a ludicrous 140.2.  Those numbers don't tell the whole story, though:

As you can see from the table, we're talking about some pretty small sample sizes for the other two guys.  They weren't put in a too many positions where they had to be in coverage, which tells us something else.  Still, those numbers are much worse than what we saw from the cornerbacks.  What about the linebackers:

Geez those numbers are bad.  It's hard to do good league comparisons -- because the Eagles did so much mixing and matching -- but just among 4-3 OLBs who played at least 25 percent of their team's snaps, Witherspoon and Gocong come out #40 and #41 out of 53 in opposing passer rating.

I mean, lots of people have criticized McDermott for all the flailing about at the end of the season (including me), but you can see what he was looking at here.  Akeem Jordan looks like the only guy who can cover anyone.  And if you stick him over the weakside, there's not much he can do about the tight end.

(Good lesson here that coaches always have a reason.  It may end up not working, but there's always a reason.)

---

So what's the bottom line?  We have to be a little careful, because we're talking about small sample sizes and all of that, but it doesn't look to me like the cornerbacks are the biggest problem they have.  You could certainly bring in a corner, move Sheldon to FS, and plug your leak on the back end that way, which seems like it would work.  And we can hope that a fully healthy Stewart Bradley takes care of that gawdawful situation in the middle for us.  But something more has to be done about defending the middle of the field.  Seems like at least one better linebacker wouldn't hurt.

Witherspoon might be the biggest puzzle.  He's not starting in the middle and he doesn't really look like he should be playing ahead of Akeem outside.  Short of unveiling the double WIL defense, I'm not exactly sure what McD can do with him.