Super Bowl Lessons

Posted on February 7, 2007 by


The big Super Bowl story is that Peyton Manning finally led his team to a victory in a big game. Such is the conventional wisdom after the Super Bowl. Unfortunately such conventional wisdom, as is often the case in this forum, falls apart when we look at the data.

The Peyton Manning Lessons

First of all, it should be emphasized that teams win championships. Players do not win titles by themselves. It’s quite possible to play well for a losing team. It’s equally possible to play badly for a winning team. And this observation is valid for all team sports.

For evidence of this in football, consider the very last paragraph added to The Wages of Wins.

Of course as we said, quarterbacks don’t win games themselves. In Super Bowl XL, the Pittsburgh Steelers won despite their quarterback, Ben Roethlisberger, who produced –2.3 net points. Matt Hasselbeck, quarterback of the losing Seattle Seahawks, produced 8.6 net points, besting Brady’s efforts from a year earlier. Again, football is a team game and it is far too easy to overemphasize the contribution of the quarterback.

Super Bowl XL was played just as we were doing the final edits of the book. Although the book was considered finished before the Steelers played the Seahawks, I thought it necessary to have one last paragraph added which highlighted how a team could win the Super Bowl despite its quarterback playing quite badly.

If we were finishing the book today, a paragraph on Manning in the Super Bowl might also be added. Such a paragraph would note that Manning’s play on Sunday offered further evidence that a team doesn’t need a quarterback to be outstanding to win the Super Bowl.

Here are Manning’s stats from the game:

Yards: 239

Plays: 40

Turnovers: 2

QB Score = Yards – 3*Plays – 50*Turnovers = 19

QB Score per play = 19/40 = 0.48

An average quarterback will post a per play mark of 1.1. So the Super Bowl MVP was actually below average in the biggest game of his career.

If we look at Manning’s performance throughout the playoffs we see that his QB Score per play was only 0.57. Given that in the regular season his per play performance stood at 3.48, Manning clearly declined quite a bit in the post-season.

Interestingly, over the previous three seasons we saw a very different story. Manning’s team kept failing in the playoffs. And people kept denigrating Manning’s game for these failures. But when we look at the numbers, Manning clearly played better each of the past three post-season.

Consider his QB Score per play during the regular season and post-season since 2003.

  • 2003: 2.94, regular season; 3.07, post-season
  • 2004: 4.37, regular season; 4.49, post-season
  • 2005: 3.18, regular season; 2.74, post-season
  • 2006: 3.48, regular season; 0.57, post-season

So what lesson has Manning learned? For his team to win, he must play bad.

No, that’s not the lesson.

There are actually two lessons. First, playoffs are a small sample and luck plays a substantial role in determining the outcome (a point made last Sunday in The New York Times). Secondly, teammates matter in football. Quarterbacks do not win or lose games all by themselves. This was true when Manning’s team failed in the playoffs. And it’s true this year as well.

So we should stop judging quarterbacks strictly in terms of whether their teams happen to win. Manning was not less of a quarterback when his team failed to win its last game. And he’s not finally a success because his team happened to win its last game.

The Rex Grossman Lesson

This lesson does not only apply to Manning. It also applies to all-time great quarterbacks like Dan Marino. And even lesser quarterbacks, like Rex Grossman.

Grossman did not play well in the Super Bowl. His QB Score per play in the playoffs prior to Sunday stood at 1.23, which is above average. On Sunday, though, he posted the following numbers:

Yards: 154

Plays: 31

Turnovers: 3

QB Score = – 89

QB Score per play = -2.87

Obviously Grossman did not have his best game.

Of course last year Roethlisberger got to be a champion, and his QB Score per play was only -1.62. So again, it would have been possible for Grossman to play badly and the Bears to take the title.

So should the Bears change quarterbacks for next season?

Often I am critical of coaches and general managers in the NBA. This is because I think the box score statistics are actually quite indicative of how productive a basketball player has been and will be in the future. So when you see decisions being made that defy these numbers (specifically decisions that are overly influenced by raw scoring numbers) I am led to the conclusion that the NBA decision-makers do not fully understand the box score statistics.

In football the numbers do not have the same quality we observe in basketball. Too much of what a player does in football depends on his teammates. Consequently we do not know if a player posts poor numbers because a) he is bad, b) his teammates are bad, c) the coaches called bad plays, d) the weather was bad, or e) All of the above.

In the end, I think football evaluation has to take place via a careful examination of the film. Only the coaches have both access to the film and knowledge of what exactly each player was supposed to do on each play. And this last part is crucial. It is very hard for an outside observer to evaluate a football player when the outsider does not know what the player was supposed to do on each play.

Nevertheless, us outsiders feel compelled to talk. So although I don’t have the same information enjoyed by the Bears coaching staff, I can still offer this observation. Grossman has only played one full season. And quarterbacks are capable of dramatically changing performance from season to season. So it’s entirely possible that a new and improved Grossman will appear next year. Of course, it’s just as possible that this won’t happen.

– DJ

Posted in: Football Stories