Padding The Stats
By Jason Lisk
How much do these garbage time yards really pad a quarterback’s stats? Well, it’s not an easy question to answer, and I don’t think I’m going to be able to do it here. Garbage time is a lot like Supreme Court Justice Potter Stewart’s “I know it when I see it” response to obscenity. It’s hard to define, and sometimes, you know it when you see it. Matt Cassel this week, for example, under any definition, would have to be considered as such.
The thing about meaningless yards is that sometimes, in rare cases, they can actually become meaningful. We don’t know until after the fact whether they mattered in the outcome, and so discounting the performance while it is occurring can be a dangerous exercise.
If we define these stat padding garbage yards very narrowly, such as down by 35, or down 20 with less than 3 minutes left, then we get extremely small sample sizes, small enough to say that it probably doesn’t matter in the overall scheme. If we define it more broadly, we get a few more cases, but probably a fair amount of false positives. After playing with some scenarios, I decided to look at it a little more broadly. Do quarterbacks pad their stats when down by more than one score in the second half? That’s a broader range, and certainly encompasses things besides just garbage time.
When I look at that scenario, I see that quarterback efficiency numbers are actually worse in these situations where the quarterback is playing from behind by 2+ scores in the second half, as the league average in these situations in 2009 was 6.01. Of course, there is a big selection bias problem with just looking at league wide numbers–bad quarterbacks and offenses tend to trail more than good ones.
To combat that, I looked at two different groups of quarterbacks. The first was the 10 quarterbacks who threw for 4,000 yards last year. The second group (since it would be useless to compare those who threw for the least yards) was those with the minimum qualify attempts (224) who had the worst qb ratings. Separately looking at these groups should eliminate the selection bias.
For the first group (the top quarterbacks), only 9% of their throws came during stat padding time in the second half. They were still slightly worse than normal during this time, as yards per attempt was down by 0.3, and everything across the board was lower, except for completion percentage. These quarterbacks were taking what the defense gave them, I suspect, which is why completion percentage was slightly up, but yards per attempt and per completion were down.
For the second group, they threw 23% of their passes during this time, and were across the board slightly worse in every category.
So, while quarterbacks throw more when trailing, and the raw numbers are increased, I don’t see any evidence that quarterback’s systematically pad their stats when trailing by multiple scores in the second half, once we account for the selection bias issue. One reason for that is that not only do worse quarterbacks tend to trail more frequently, but better pass defenses tend to lead more frequently. It becomes easier to put up good numbers when you happen to play a bad pass defense when trailing.
This week, we had a perfect storm of several teams jumping out to a lead and playing well in the first quarter, who have not had very good pass defenses this year. Predictably, then, these defenses gave up yards as the game unfolded, when the other team passed heavily. Sometimes you know it when you see it, and this week saw it’s share of garbage time yards. I don’t see any reason to universally discount quarterbacks, though, who play well when their team is trailing. [photo via Getty]