The NFL Playoffs and Steroids in Baseball

For the past few weeks I have been posting less and working on the second book more (really).  In the middle of Saturday afternoon, though, I posted a comment on Andray Blatche.  And on Sunday morning I am here commenting on the NFL playoffs (and steroids in baseball).  So what’s going on?

This weekend I can’t work on the book.  I have a bit of grading to do, and so the book has to be put aside.  Of course there’s a problem.  Grading is not my favorite thing to do in the world.  Consequently, I have spent my weekend playing with data and doing a bit of writing on this blog (later today I will do all my grading…. really).

The NFL Playoffs

After playing with NBA data (and talking about Blatche) my thoughts turned to the NFL.  Specifically, I decided to download playoff data from the NFL and see if I could offer any amazing insights on the 2007 post-season.

Table One: QB Score and RB Score Rankings for the 2007 Playoffs

Table One offers an analysis of the quarterbacks and running backs who received significant playing time (14 pass attempts per game for quarterbacks, 10 rushing attempts per game for running backs) in the post-season.

In terms of the NFL’s QB Rating, the top quarterback in the post-season has been Tom Brady.   When we turn to QB Score per play, though, Brady only ranks 5^th.  Phillip Rivers, Peyton Manning, Eli Manning, and Brett Favre all produced better stats per play than Brady in the playoffs.

Turning to the running backs, we see the top back in the playoffs has been New England’s Laurence Maroney.  Second on the list is Marion Barber.  Both players attended the University of Minnesota.  This probably doesn’t tell us much, but is probably interesting for Golden Gopher fans (or not).

Looking past the Golden Gopher comment, do we learn anything from Table One? When it comes to football I have a few insights.

1. QB Score, relative to the NFL’s QB Rating, is more accurate, more complete, and quite a bit simpler to calculate.

2. RB Score is also more complete (and accurate) than simply rushing yards per game.

3. Regardless of metrics used, though, the statistics we see from NFL players are very inconsistent.  So what you see in the past is no guarantee of what you see in the future.

4. The inconsistency in player statistics is primarily due to the interdependency between the player and his teammates.

5. Even if this interdependency issue was not much of an issue in football – or in other words, even if football was like basketball – quarterbacks and running backs only play half the game.  So the credit and blame given to these players often overstates their contribution.

When we look at this list of insights, what do the above rankings tell us about the Super Bowl?  Hmmmm… not much.  Past performance can’t really predict the future very well. In other words, I don’t think anyone could have predicted how well Eli Manning was going to play in the playoffs.  Eli truly was a below average quarterback in 2007.

Furthermore, even if we could predict the play of Manning and Brady, the defenses of New England and New York (as well as the special teams) will have a significant outcome on the Super Bowl. 

Better Football Columns

All in all, this is not too satisfying of a column.  Fortunately there are a few columns I have seen recently that I think do provide more satisfying insights.

Carl Bialik, the Numbers Guy (a Wall Street Journal blog), offered a few thought on the NFL’s QB Rating.

The NFL’s Most Mysterious Number

Bialik’s column does a nice job of reviewing the problems with the NFL’s metric.  Plus it mentions QB Score, so that alone makes it a good read (for me at least).

Bialik doesn’t normally comment on football.  That cannot be said for Brian Burke and Doug Drinen.  Burke’s blog – NFL Stats – has two posts that I think will be of interest. The first offers a forecast of the Super Bowl.  This post gives us an estimate of New York’s chances are in this game.

Super XLII Prediction

The second Burke post examines whether quarterbacks are capable of performing differently in the Red Zone.  I think his analysis will be surprising for some.

Is “Red Zone Performance” Real?

Drinen is a math professor at the University of the South (and hence a former colleague of JC Bradbury of Sabernomics).  Drinen, like Burke, routinely offers very good columns on professional football at the pro-football-reference blog.  I am going to provide links to two of his latest.

The TD Project

Approximate value II

The second of these is especially interesting.  In this column Drinen tries to construct a measure that allows us to compare quarterbacks, running backs, offensive lineman, etc…  This is definitely a difficult issue to tackle (pun intended) and Drinen’s efforts should provoke some thought.

Steroids and Baseball

Let me close with a comment on a completely different subject.  Steroids and baseball has become a hot topic in the sports community lately.  I think the aforementioned JC Bradbury is “the expert” on this topic.  I highly recommend you click on the following link, which takes you to all of his writings on steroids at Sabernomics.

JC Bradbury on Steroids in Baseball

And for more on this topic, please see Alan Schwarz’s latest column in the New York Times.

A Voice of Skepticism on the Impact of Steroids

Schwarz primarily focuses on the work of Eric Walker and Walker’s website: Steroids-and-Baseball.com.  Walker argues that what people generally think about the impact steroids have on baseball is incorrect. 

One final note… what Walker says is consistent with what Bradbury has said. And yes, Schwarz also quotes Bradbury in the column.

– DJ

For more on QB Score, RB Score and what these metrics mean see

The New QB Score

Consistent Inconsistency in Football

Football Outsiders and QB Score

The Value of Player Statistics in the NFL

Advertisement

Posted in: Football Stories