Taking out the garbage...That is exactly what I propose to do with this article: take out everyone's garbage time stats and see what's left.

First we need a definition of garbage time. I'll define a situation as garbage time if any of the following conditions holds:

1. there are 5 minutes or less remaining in the game and the team with the ball is trailing by more than 8
2. there are 10 minutes or less remaining in the game and the team with the ball is trailing by more than 10
3. it is the second half and the team with the ball is trailing by more than 14

These situations are probably better described as "loose defense" or "playing catch up" rather than garbage time, but I rarely stumble across a clever title, so I took some liberties. The point is that conventional wisdom tells us that teams trailing by a substantial amount are forced to pass more often, thus leading to better numbers for quarterbacks and receivers. It also tells us that passing is actually easier in the situation, as defenses are more willing to give up some cheap yards in the name of guarding against the quick score.

This would then suggest that, all else equal, we might prefer a receiver on a team with a bad defense to one on a team with a good defense. It also might point out some players whose stats were inflated by garbage time last year. To take one example, consider Anquan Boldin. If it really is easier to pass in garbage time, and if the Cardinals really did have a lot of garbage time last year, then Boldin may not be as good as his 2003 numbers appear.

Those are two big ifs, though. Is it easier to pass in garbage time? Did the Cardinals have a lot of garbage time last year? Well, OK, that second one isn't such a big if. Twenty-seven percent of the Cardinals' snaps last year were in garbage time, which led the league. Here's the full list:

PERCENTAGE OF OFFENSIVE SNAPS THAT WERE IN GARBAGE TIME

So how much did all that garbage time help Boldin? Quite a lot, it appears. Consider the following numbers:

Boldin was more than twice as productive in garbage time. So let's take out the garbage. The Cardinals ran 928 offensive plays last year. If we pretend that all of them were non-garbage and if we assume --- and this is a questionable assumption, which we'll discuss shortly --- Boldin produced at his non-garbage fantasy points per snap rate, then Boldin would have had 139 fantasy points. That's 48 fewer than he actually had! A case could be made that Arizona's generous endowment of garbage time added 48 fantasy points to Boldin's 2003 total.

Two questions immediately arise:

1. How reasonable is that assumption we made above? Namely, is it fair to assume that Boldin's non-garbage points per snap average is more representative of his true abilities than his overall average is?
2. How does Boldin's figure compare to other wide receivers? We estimated above that garbage time inflated Boldin's numbers by almost 50 fantasy points last year. But if all the other top receivers also gained roughly as much, then there's no reason to downgrade Boldin's numbers.

Let's answer the second question first. To do that, we'll look at all receivers who scored at least 100 total fantasy points last year and run the same calculations that we ran above for Boldin. Namely, the number in the second column below is how many points garbage time added to the player's stats in 2003:

So Boldin's case was an extreme one. He benefited from the garbage time to a far greater extent than anyone else. Note that a lot of his colleagues near the top of the list are, unlike Boldin, players who capitalized on a very limited amount of garbage time.

Also worth noting is that several big-name receivers actually played worse in garbage time, so this calculation indicates that garbage time actually robbed them of some points last year. Particularly interesting are the numbers of Randy Moss, who has a reputation for quitting when he doesn't feel the game is winnable. This data, while certainly not conclusive, does provide some additional fuel for that fire.

But none of this amounts to anything worthwhile until we've answered the first question: is it fair to assume that these receivers' splits are indicative of a real effect? Or are they generated by random chance?

Let's take a look at all wide receivers who had 100 total fantasy points or more last year and list their points per team snap average in both garbage and non-garbage situations.

An interesting statistical phenomenon is on display here. The non-garbage rates are fairly tightly bunched, but the garbage rates vary widely. This is to be expected, because the garbage rates are based on a much smaller sample size. Remember, about 11% of all plays were run in garbage time league-wide A glance at the top of the list indicates that some very big name receivers --- Holt, Mason, Harrison, etc. --- feasted on garbage time last season. It is very interesting to note, however, that Boldin is the only receiver with a drastic split who played a significant amount of garbage time. Harrison, Mason, Holt, Steve Smith, and Javon Walker all came from teams that played a minuscule percentage of their snaps in garbage time, and their garbage numbers are therefore based on extremely small samples.

Overall, 18 receivers were better in garbage time and 15 were worse. This suggests that there might be a slight tendency for receivers --- on the whole --- to post better numbers in garbage time. But Boldin and, to a lesser extent, Peerless Price, were not slightly better in garbage time. They were much better. They were different from the norm, and so they deserve some extra examination.

Random chance can make splits materialize even when no real effect is present. The Cardinals played 27% of their snaps in garbage time. If you took 27% of the Cardinals snaps at random and looked at Boldin's numbers in that 27%, it is possible that you would find a split as big as this one. How possible? Well, there are some fancy statistical techniques that can answer that question, but they ultimately do not and cannot answer the question of whether Boldin and Peerless Price really did have their numbers inflated by the garbage time. In fact, it's not stretching the point to say that we will never know the answer. All we can do is speculate.

Personally, I think that there may be something real here. Normally, I scoff at the notion that breakout players achieve their breakouts because defensive coordinators are not aware of or concerned about them. But Boldin was a rookie last year, and it is easy to imagine some combination of "I can't believe he's really that good" and "even if he does rack up some yards, we've still got a comfortable lead" attitudes combining to create an easy atmosphere for Boldin to compile some cheap stats. Likewise, Peerless Price has a reputation as a guy that is great if he has help. In Buffalo, he had help in the form of Eric Moulds. Maybe last year in Atlanta, he had help in the form of halfhearted garbage time defense.

On the other end of the spectrum, the cases of Randy Moss and Laveranues Coles are fascinating. Minnesota had an above average amount of garbage time last season, and Moss performed very poorly (for him) during that time. Will an improved Viking defense cause Randy to have his head in the game a greater percentage of the time? On some level, it would be absurd to expect Moss to improve upon his 2003 numbers --- you just don't have a once-in-a-decade type season and then improve on it --- but these numbers suggest a hint of a possibility that there is room for just that. Laveranues Coles did not perform well in garbage time either. Why not? I cannot think of a reasonable explanation, which means that there may not be one. But if there is, a generally improved Redskin team should cause his numbers to head north.

In my next article, I'll take a look at running backs. Does garbage time help or hurt LaDainian Tomlinson? That is not a question with an obvious answer. I'll also examine the other side of this equation: clock-killing time. In particular, which runners were helped the most (and which receivers harmed the most) by the fact that their teams were frequently milking leads at the end of games?