A Tale of Two Seasons?

One of the things every Eagles fan knows is that Donovan McNabb struggled mightily in the first half of last season, but rebounded nicely in the second half after he had some time to better recover from the catastrophic knee injury of 2006 (not to be confused with the catastrophic sports hernia of 2005 or the semi-catastrophic broken ankle of 2002). 

Just check out the per-game stats (Miami removed because of the injury):

Pretty black and white, right? 

Nah.  Never that easy. 

Here's the problem:

McNabb actually played a great game in week nine against the Washington Redskins.  It's the inclusion of those results that so dramatically skews the first half / second half results.  Here's that chart from above if you count weeks 1-9 and then 13-16:

Certainly some improvement, right?  But not the massive performance increase everyone keeps talking about.

And I would argue that this way of breaking down the season makes far more sense.  If you want to split it after week eight and use the "injury comeback" explanation, you have to explain how McNabb magically healed so much between Dallas and Washington that he was able to drop four TDs and a 138.5 passer rating on a decent pass defense.

I'm open to suggestions, but that's not going to be an easy story to tell.

Now I guarantee at least some people right now are saying to themselves: "Yeah, but if you took out the Detroit game, that wouldn't look so good."  And that's certainly true.  But why is it only the Detroit game that doesn't count?  Shouldn't we be just as ready to strip out the season-ender against Buffalo?  After all, it was a meaningless game between two non-playoff teams.  Most guys were planning their vacations by that point.

Here's another thing to consider.  Check out a graph that matches McNabb's passer rating (red line, left axis) with the rank of the pass defense he was facing (blue bars, right axis):

He should have looked good those last two weeks -- he was facing two of the worst pass defenses in the league.

Now I'm not trying to argue that McNabb didn't get better as the season progressed.  He absolutely did.  But he wasn't all that bad in the first half of the year either.

I think people feel like McNabb made a big jump in the second half for three reasons: 

1. He ran more.
2. No one sacked him 12 times.
3. The Eagles won four of the five games he started and finished.

But those last two reasons have as much to do with the rest of the team as they do with him.  After all, if Kevin Curtis hadn't busted his butt downfield to recover McNabb's fumble in the first quarter against the Saints, we might not be talking about "just" an 8-8 team.

In the end, team wins and losses dramatically skew our perceptions of individual performances.  Heck, all anyone wants to talk about these days is the emergence of Eli Manning, like last year was the season he finally put everything together.  Well, the guy's season-long passer rating was 73.9.

Or more than 12 points worse than McNabb's "crummy" first eight weeks.

Comments

[Zero]

Great analysis, IB but stats don't always tell the whole story. Just watching DMac last year one would notice a big difference in his body language from the first half to the second. I don't know if it was his confidence, the brace, his leg strength or what, but he sure looked a lot better towards the end of the season from my perspective. Last year was an anomaly: McNabb, Kearse, LJ injuries, Reid family issues, complacency on the OL all came together at the same time.

[Dave B from OC]

Wow, great entry. Super 5 is always going to get blamed if the Birds lose, regardless of how he plays. I thought he played well enough, individually, to win most weeks last year. Can't say the same about the team though. They seemed distracted or just unprepared more weeks than not.

[taco pal]

Just my second visit, but I think you've got a really fantastic blog.

That said: it strikes me that the Detroit game was a rather extreme outlier, and that if, let's say, you were to use median stats instead of mean stats, the result you'd get would much more closely reflect the conventional wisdom. Or at least that's what it seems like to me, just from eyeballing it.

[Me]

The passer rating statistics are calculated from the raw numbers, they aren't averages.

I just ran the other numbers with medians, rather than means, for the 1-9/13-16 weeks, and got this:

Cmp Att Yd TD Int Sck
21 34 251 1.0 0.0 3.0
24 38 236 1.0 0.0 3.5

I think the big thing with the eyeballing factor is that McNabb definitely had a slow first month, but after that he was pretty much on track.

I don't mean to suggest that McNabb wasn't better in the second half of the year. He clearly was. It's just that it's not as cut-and-dried as a lot of people are making it out to be and our perceptions are somewhat skewed by the late season wins. We feel like he *must* have been better because they won those games.

And welcome to the blog.