A few days ago I commented on the performance of NFL quarterbacks in the playoffs this season. The results were less than satisfying, since my data set only included 10 playoff games.
Fortunately Yahoo! Sports reports complete playoff statistics for quarterbacks back to 2001. With this larger data set we can address two issues:
Is there a relationship between playoff performance and what we see in the regular season?
Are there any quarterbacks who consistently perform better in the playoffs?
Playoff Consistency
To answer the first question I collected regular season and playoff data on every quarterback since 2001 who met the following thresholds:
- Attempted at least 28 passes in the playoffs.
- Attempted at least 224 passes in the corresponding regular season.
Since 2001 there are 32 quarterbacks who have met these standards once, and since several have done this multiple times, the data set I constructed has 63 total observations. I then ran a simple regression of each quarterback’s QB Score per play in the playoffs in the regular season on what he did in the corresponding regular season per play. The results indicate that there is a statistically significant relationship. But the explanatory power of the model is quite low. Only 9% of what a quarterback does in the playoffs can be explained by what he did in the regular season. [Interestingly, when I raised the threshold for entry into the data set to 50 plays in the playoffs – a restriction that left me with only 40 observations – the explanatory power of the model dropped to 6%.]
To put this result in perspective, consider what we reported for the NBA. In The Wages of Wins (and in a research note I published with Erick Eschker in The Journal of Economic Issues) we reported the relationship between what an NBA player does per-minute in the playoffs (measured with Win Score) and what he did in the corresponding regular season. Although we considered a number of factors in both the book and the article, when you do the simple regression of playoff performance on regular season performance you find that 63% of what an NBA player does on the game’s biggest stage is explained by what he did in the regular season. In other words, as we find when we look at performance in the NBA from season-to-season, we also find when we look at the post-season: Relative to what we see in football, basketball players are fairly consistent across time.
Of course, NFL quarterbacks – relative to what we see in basketball – are really inconsistent. And this is not just true in the regular season (which I have noted several times in this forum, but perhaps was best stated HERE) but also when we look at the playoffs.
The Prime Time Performer
So predicting playoff performance is difficult. We stilll have the second question.
Is there any quarterback who consistently played better in the playoffs? Much to my surprise, there was one quarterback who has appeared in at least three post-seasons since 2001 and raised his play in the playoffs every single time.
Before I get to the name of that player, let me list the quarterbacks who have made at least three playoff appearances since 2001 (number of appearances in parenthesis): Tom Brady (5), Brett Favre (4), Matt Hasselbeck (4), Peyton Manning (5), Donovan McNabb (4), Steve McNair (3), Chad Pennington (3), and Jake Plummer (3). On average these eight quarterbacks posted a 2.07 QB Score per play in the corresponding regular season (average is 1.1) and in the playoffs saw there per play performance dip to 1.33 (as we note in The Wages of Wins, part of the reason playoff performance dips in the NBA is because the quality of the opposing team increases. A similar phenomenon is likely going on here).
The biggest improvement we see with these seven quarterbacks was posted by Brett Favre, who posted a 1.11 QB Score per play in 2003 and then raised his per play performance to 3.41 in the playoffs. Hence his per play performance increased by 2.29. Interestingly, in the other three times Favre appeared in the data set his per play performance declined every time. And the magnitude of the decline was always greater than the increase we see in 2003. In 2001 he declined by 2.67, in 2002 it was a decline of 2.44, and then in 2004 his per play performance dropped 5.29. The decline we see in 2004 was the largest drop-off by any quarterback since 2001.
Favre, though, is not the only one to suffer such a large drop-off in performance. Per play declines of two or more were posted by Jake Plummer (twice), Peyton Manning (twice), Steve McNair (twice), Chad Pennington, and Donovan McNabb. For Manning, one of his major declines was observed this year. In 2006 Manning has so far dropped 2.88 from the per play number we saw in the regular season. Despite this, he is still playing in the Super Bowl next week.
Okay, which quarterback has never dropped off in the playoffs (at least, not since 2001)? One might suspect Tom Brady. Brady did play better in the playoffs in 2001, 2003, 2004, and 2005. But in 2006 his per play numbers in the playoffs were worse than what he offered in the regular season.
Okay, if you are counting, there is only one name left from the original list of eight quarterbacks. Yes, our prime time quarterback is Matt Hasselbeck.
Hasselbeck is somewhat famous for saying in an overtime playoff game in 2004: “We’ll take the ball and we’re gonna score.” After making this pronouncement, Hasselbeck proceeded to throw an interception that was returned for a touchdown, giving Brett Favre and the Packers the playoff victory. Despite this setback, Hasselbeck’s playoff QB Score per play — in four post-season appearances – has always bested his corresponding regular season performance.
Here are the records of the seven quarterbacks in our sample:
- Matt Hasselbeck: Improved four times, Never declined
- Tom Brady: Improved four times, Once declined
- Peyton Manning: Improved twice, Three times declined
- Steve McNair: Improved once, Twice declined
- Chad Pennington: Improved once, Twice declined
- Jake Plummer: Improved once, Twice declined
- Brett Favre: Improved once, Three times declined
- Donovan McNabb: Never improved, Four times declined
McNabb is the only quarterback in our sample of eight to never play better in the playoffs. McNabb did appear in the playoffs in 2000. And looking at the box scores reported in Sporting New 2001 Pro Football Guide, it appears that McNabb was also worse in the playoffs in 2000 (although the Pro Football Guide does not report fumbles lost). So McNabb has yet to play better in the playoffs. One should note, though, that McNabb’s declines are typically small.
Just the Numbers
Although McNabb is the only one of these eight to never improve in the playoffs, his average post-season performance is not the worse. Here are the QB Score per play for these signal callers in the playoffs and regular season since 2001:
- Matt Hasselbeck: 2.25 (playoffs), 1.77 (regular season)
- Peyton Manning: 2.08 (playoffs), 3.13 (regular season)
- Tom Brady: 1.89 (playoffs), 1.62 (regular season)
- Donovan McNabb: 0.99 (playoffs), 1.82 (regular season)
- Chad Pennington: 0.79 (playoffs), 2.01 (regular season)
- Steve McNair: 0.71 (playoffs), 2.03 (regular season)
- Brett Favre: 0.03 (playoffs), 1.68 (regular season)
- Jake Plummer: -0.01 (playoffs), 2.39 (regular season)
As we can see, all of these quarterbacks were above average in the regular season, but only three were above average in the post-season. Hasselbeck is the top playoff performer in the post-season, following by Manning and Brady. Interestingly, Manning – on average – has actually performed better than Brady.
The Lessons Learned
Unlike my previous post on NFL quarterbacks in the playoffs, I think this study tells us something. First, there is very little relationship between what we see in the playoffs and regular season. So those who are trying to guess what we are going to see in the Super Bowl, keep in mind that all you are doing is guessing. We also find that quarterbacks typically play worse in the playoffs. Finally, it appears one quarterback has managed to defy the tendency to play worse.
Does this mean that Hasselbeck has some kind magical playoff formula? Well his team has yet to win a Super Bowl, so it’s not exactly the case that he (and/or his teammates) is playing perfectly in the playoffs. My guess is that so far Hasselbeck has been a bit luckier than the average quarterback. And if he keeps appearing in the post-season we will eventually see him play worse.
Let me close by asking if anyone has playoff data for quarterbacks prior to 2001 – and I mean passing yards, rushing yards, yards lost from sacks, passing attempts, rushing attempts, sacks, turnovers, and fumbles lost – please send it along. I have been able to find much of this, but fumbles lost remains elusive.
– DJ
QB Score has been discussed previously in the following posts:
Football Outsiders and QB Score
Consistent Inconsistency in Football
The Value of Player Statistics in the NFL
Simple Models of Player Performance
Wages of Wins Journal 
Posted on January 28, 2007 by dberri
0