Love him, hate him, think he can bring us a championship next year, think he never will — opinions on McNabb run the gamut. Yet when we consider the arguments for keeping or trading him, we have to eliminate those biases and try to approach the decision from a rational perspective.
Is it unwise to trade a proven QB? What if you've been grooming his successor for years? Will a package of picks for McNabb make up for the loss of elite play from McNabb? Is McNabb actually elite?
Of the questions, today I'm going to examine the last one: how good is Donovan really?
Derek over at Igglesblog as already given us one look at this in Was This McNabb's Second-Best Season?
Derek realized, correctly, that you can't always judge a player's performance against his past numbers. If the league is getting more pass-friendly, McNabb's stats have to be compared to other quarterbacks to see his true worth. He puts together a great chart of statistics (INT%, YPA, Y/G, QB Rating, TD/G) on how McNabb ranked each year in those categories. I put those numbers into a graph so we could more easily see his performance.
Looks pretty good overall. You can see the blue is McNabb's year to year average rank in these five categories. The black is just a trendline to follow more easily. I also threw Kevin Kolb's 2.5 game numbers into there for 2009 (obviously small sample size caveats apply) just so you can get a sense of what he might have ranked if he continued to play at that level.
This jives with most of our conventional wisdom: that Donovan has been a top 10 QB since at least 2004. As Derek noted, maybe not top 5 anymore, but consistently around 10th best the last 3 years. That seems like performance that, if it can be counted on, should not be abandoned lightly.
But do conventional stats tell the whole story? What about Football Outsider's DVOA, which is supposed to more accurately show player performance over the average.
Again, don't read too much into the Kolb number off 2.5 games. But DVOA has similar results to the conventional stats. McNabb's been at least 10% more valuable than the average QB since 2004 , with the last few years' average oscillating around 13%.
There is obviously a good deal of variation. Ultimately any given year's performance will fall across a range. I've tried, with the trendline, to show the general change without big jumps year to year, and I think it does a pretty good job (considering the changes from year to year in the Eagles offense) isolating McNabb's performance. We're probably looking at somewhere between 10-15% above-average QB play if McNabb can keep it up.
Unfortunately, with DVOA we run into the same problem as all the other regular stats above. It's based on a multi-season average and over the last couple years QBs have gotten better and better according to it. So where does McNabb stand up according to his peers?
Again, rather large year to year jumps. But overall, the trendline isn't nearly as constant for McNabb's recent performance. Examining his play versus other QBs shows that he hasn't been a top 10 QB since 2006.
The peak from 2004-2006 corresponds to the peak in the other graphs — but it isn't nearly as high. According to this measurement, McNabb has never been a top 5 QB. And rather than a consistent top 10 QB, the past three years have shown him averaging around 15-16th among his peers.
At the end of the day I think this final graph gives us the clearest picture of what we're dealing with in McNabb. He is a good quarterback, but not elite. At this point he is not providing much better than average performance compared to the rest of the league.
Yes, we can try to come up with theories to explain away the decline in DVOA rank: lack of #1 wide receiver, lack of running game, etc. But each of these recent teams were very different in their flaws. 2009 had a weak running game and young WRs. But 2008 had worse wideouts. 2007 had a better running game.
And no, this does not mean McNabb is bad. He is absolutely capable of leading the Eagles to a Super Bowl win — worse QBs have done so. And give him the right weapons or playcalling or situation, and maybe his DVOA could spike from its current general position.
But these numbers suggest on the whole that his own play is either stagnant or declining (depending on how you analyze the graph and the future), while quarterbacks on the whole are getting much better. McNabb's relative value is rapidly decreasing.
The question then becomes: what is it worth to keep a steady/declining, injury-prone, average QB starter? Could Kolb provide average QB play for youth, less money, and more upside?
If you think Kolb can be at least "average" — a decision based on limited starting experience — then the choice is clear. You take the 26 year-old (hopefully ascending player) over the 33 year-old (likely descending player) every time.
Follow me at www.mcnabborkolb.tumblr.com — come get all your latest news and analysis on the Eagles Offseason QB Clusterf%$k™