Closer Look at the Red Zone

I've been mulling over another red zone post for some time now.  I covered it a couple times during the season, but not in a really systematic way with year-end stats.  So that's what I'm hoping to do here.

First, a quick aside.  I started working on this post over the weekend -- these stat-heavy ones take awhile to work through -- and didn't see that some good stat work had been taking place on the same issue over at Wingheads.  Since that's out there, I figured I'd better address it and then move on to a broader picture.

Here's Ig_L's thesis:

I broke the season down into halves and took out the Detroit blowout. Well we were worse than pathetic in games 1-8 (excluding Detroit). We scored touchdowns on 26.1% of our Red Zone appearances...

In the 2nd half of the season we really improved. We had a 61% TD rate (NFL average is 52%). This would have put us in 4th place in TD percentage...

What does this mean? At a minimum, it means we CAN be in the top 5 in Red Zone production. I don't know what the cause of the problems in the first half of the year, but they were eliminated in the first few games of the 2nd half. Then we slid back to just above average in the last few games.

He's done up a number of good tables to show his work (although his numbers differ from the ones I've found, inconsistently).  But I don't find this convincing.  You can't argue for a trend that says:  "We were bad for two games, then good for one, then had six bad ones, four good ones and then a few OK games."  That's not a trend, that's just normal variation. 

Furthermore, there are just too few red zone opportunities on a week-to-week basis to say anything meaningful about performance from one game to the next.  In fact, you can take that idea another step further and argue that ANY attempt to understand red zone performance is meaningless due to the unavoidably small sample sizes (unless you're the Patriots). 

This is what Brian Burke, who runs the football stats site www.bbnflstats.com, told me when I emailed him back in early March to get his take on the issue.  Brian is clearly a very, very smart guy.  And the stuff he does with statistics is way beyond what I usually cook up.  You should definitely check out his site when you have some free brain cells. 

His response to my email:

From the analysis I've seen and done, there is nothing special about any offense that makes them particularly well suited or poorly suited to red zone success. In other words, teams that are good elsewhere on the field will tend to be good near the end zone, and teams that are generally bad will tend to be bad near the end zone.

Red zone performance is a small subset of general offensive performance. It is likely that it is no more different from overall performance than if you took any random subsample of all plays. For example, say that red zone plays comprise 5% of a team's offensive snaps. If you took any 5% sample of plays from a random location on the field and looked at performance there, you would see the about the same deviation from overall performance.

There probably aren't enough plays in the red zone to compare to overall performance to conclude that any team is significantly better or worse in the red zone than anywhere else. And if there were enough plays to analyze, the effect would necessarily be so small as to not be that important. It's a little bit like "clutch batting" in baseball. Statistics can't prove it doesn't exist, but it can conclude that if it did, the effect would be tiny.

So, I would take heart with the Eagles. Their very high rate of reaching the red zone indicates they have the talent to move the ball. Their lack of red zone TDs last year are probably due to 1) randomness, and 2) game situation (being either ahead or in close games so that FGs are attractive).

I found some data to support what he's saying, too, at ProFootballWeekly, another good statistical site that tracks red zone performance throughout the season.  Check out the following graph that shows red zone scoring efficiency for every team in the league in 2006 and 2007:

If red zone scoring were a "skill" we would expect teams that were good at it one year to also be good at it the following year, since there isn't that much turnover in rosters, coaching, etc., from year to year. 

But that's not what we see with this chart.  Rather than the dots "clumping" around an imaginary line that rises from left to right, we actually have what appears to be a rather random distribution.  (Correlation coefficient of -0.09, in fact).   

The situation isn't quite the same for touchdowns, but it's pretty close:

There's more of the clumping here (correlation coefficient of .23) but it's still not a huge effect.  By the way, that data point over on the left is the Oakland Raiders. 

So case closed ... or not.  Take a look at the Eagles' five-year trends in terms of red zone performance:

Looks can certainly be deceiving, but does that look random to you?  Because it kind of doesn't to me.

- - - - - -

So why was the team so good at scoring in the red zone in 2003?  Or, to rephrase that for our stats-minded friends, in what way did the Eagles score all those touchdowns that year?  Here's an interesting chart:

That's pretty intriguing, isn't it?  In the last five years, you have three seasons with 15 red zone passing TDs, one with 14 and one with 27. 

Here's the full table:

I find this almost insanely interesting, because people have a tendency to say, "Well, the Eagles used to be just fine scoring in the red zone when they had Todd Pinkston and James Thrash, so clearly they should be OK now."

Uh no, the reason the Eagles were so good at scoring touchdowns in 2003 is because the Three-Headed Monster of Westbrook, Buckhalter and Duuuuuuce was so good at pounding the ball into the end zone on the ground.

So could that be the answer?  Just run the ball down more by the goal line and watch the points pile up?

One problem.

- - - - - -

For the next section, I'm just going to steal from an earlier post I wrote.  I haven't updated the numbers (life is too short), but I continue to think this is really a key point for the Eagles:

And what we saw this year is that Westbrook is too good to be shut down in the middle of the field.  The coaches can do too many things to get him the ball in space for any team to consistently hold him in check.  But when the Eagles get down by the goal line, things change.  The space compresses, safeties who no longer have to worry about helping deep can start creeping up, and there's just not as much room to spring Brian free. 

I put together a couple of tables to show just how big an effect this has.  The first compares Westbrook to the other backs in the league's top 10 in rushing:

Now obviously, all these guys are going to have lower averages in the red zone, just because there's not as far to run when you're close to the goal line.  But Westbrook has the biggest gap of all these guys between his rushing average in the red zone vs. the rest of the field...

In short, I think a big part of the reason the Eagles have had a lot of success between the 20s and not as much in the red zone is because you can't defend against Westbrook in the first case, but you can in the second.

I still think that's true.

- - - - - -

But what about 2004, anyway?  If we compare 2004 (Owens) to 2007 (no Owens), what will we find?

Weird, isn't it?  Inasmuch as Owens was a bona fide red zone threat, it wasn't really just about him.  Owens, Westbrook and LJ all had six red zone TD catches, with Chad Lewis chipping in five more.  In 2007, no one guy had more than three.

In fact, if you just look at the wide receiver position as a whole, the group managed nine touchdowns in 2004 ... and nine touchdowns in 2007.  The difference was all the other guys (especially the tight ends).  Did Owens really draw so much attention down by the goal line that it could have that big an impact on everyone else?  Or was the trio of Smith / Lewis / Bartrum just that much better than Smith (injured) / Celek (rookie) / Schobel (sucks)?  And how much goes on the shoulders of the rehabbing McNabb?

Not sure there's any way to tell.

- - - - - -

So what's the lesson here?  That it's not about having that one guy who's your "go to" red zone threat?  But, rather, what matters is being able to spread the ball around and keep the defense off balance so that everyone has better chances to score?  Or does it take that big time guy to open things up for everyone else?

And would spreading the defense out mean that the running backs would find more room to run?  Maybe increasing scoring chances even without throwing the ball?

Most importantly -- have the Eagles done enough this offseason with the additions they've made to get to that "spread out" level? 

I guess we'll see.

Comments

[Lex]

Near the end,

In 2004, no one guy had more than three.

Should be:

In 2007, no one guy had more than three.

Great post!

[Me]

Thanks for the catch.

[dmac]

Yeah, really interesting post.

[flee]

excellent post. the nfl stats site you linked is awesome. and i agree with him, i don't see any systematic reason why the eagles weren't good in the red zone last year. it was just a bad year. honestly, if i never hear the words "we need a #1 receiver to help in the red zone" again, i will be happy. this whole season is riding on the health of #5 and #36. if they are healthy the eagles will be super bowl contenders, if not, it will be another long year.

[Mike in Vegas]

Did you catch what Andy Reed said the other day in the interview with WIP that Jason posted? "Andy blames the lack of production in the red zone on the lack of TE play they had rather than WR play."

Looks like your stats back that up.

[rage114]

Great article. I was about to do some research on this very topic. You did a much better job than I would have and I appreciate the effort.

[Ig_l]

The only things I don't agree with are:

.23 correlation coefficient may not be all that weak considering the changes in teams fortunes from one year to the next (see Eagles 2004 vs. 2005).

Regarding no (significant) difference between performance in RZ vs. the rest of the field: because of the limited space available in the RZ and end zone, there are some differences in both offensive and defensive plays and the personnel they use. For example, I'd expect more TEs and bigger backs in the game, especially inside the 5 yd line.

[Me]

I actually agree totally. But getting into a discussion about clutch hitting / red zone performance / streak shooting with the statistical people is somewhere just behind "never get involved in a land war in Asia" on the list of things I try to avoid.

With that said, you have to admit that even with the Eagles, it's a little weird. I mean, they were very good in both 2003 and 2004 in the red zone, but they couldn't have been more different in _how_ they went about it.

I really should run all the year's data that I have to see if that .23 holds up.

[Eagle Man in San Fran]

Great post (and blog)!

I think it IS significant that the Eagles were much better in the red zone the second half of last year - and that one can mainly attribute it to McNabb finally recovering from his ACL injury.

Last year, I did some number-crunching for the first 6 games of the year (excluding the Lions game) for the Eagles in the “red zone”. Two things stand out:
1) Pass/Run Ratio: 49 passes, 21 runs (with 2 being by McNabb) or a 70% pass ratio
2) Pass Stats: 19/49 (38%) for 111 yards (2.2 yards/attempt)

So there you had AR saying to McNabb, “You get us into the end zone” and McNabb couldn’t do it, as reflected in his horrible passing stats.

But I would contend that McNabb’s lousy passing stats weren’t due to poor passes (evidenced by the fact that he wasn’t intercepted in the red zone at all in those 6 games), but rather his lack of being able to make a play with his feet, plus having less time to throw (due to the injury not allowing him to “buy more time” with his feet).

Thus, I’d be interested to know what the stats were for the second half of the year (sorry, but don’t have time to do it myself). The pass ratio in the red zone may not have been different, but I would think McNabb’s red zone passing stats were much, much better the second half.