Writers of Pro Football Prospectus 2008

22 Aug 2007

Letting T.O. off the Hook

When I did Bill Simmons' "BS Report" podcast, he suggested four different stats he wanted us to calculate. One of them would measure "killer drops on third downs." Simmons called it the "T.O. number,"speculating that Owens led the league in this stat for 2006.

I was pretty sure that Owens led the league in this stat too. We were both wrong. Here are the 2006 leaders in drops on third or fourth down, according to our game charting project:

8: Torry Holt
7: Troy Williamson
5: Drew Bennett, Alge Crumpler, Vernon Davis, Randy Moss, Dancin' Reggie Williams
4: Joey Galloway, Bryant Johnson, Eddie Kennison, Randy McMichael, Terrell Owens, Jerramy Stevens, Bobby Wade

T.O. led the league with 15 drops, but very few of them were on third down. He had three on first down, eight on second down, three on third down, and one on fourth down. One of those third down drops wasn't even near the sticks -- it was a two-yard pass on third-and-7 against the Giants.

On the other hand, eight of Holt's 10 drops (tied for sixth in the NFL) were on third down. Six of them were sure first downs, and another one was a five-yard pass on third-and-7, which probably would have moved the sticks thanks to YAC.

By the way, here are the 2005 leaders in drops on third/fourth down:

8: Muhsin Muhammad
5: Antonio Bryant, Plaxico Burress
4: Terry Glenn, Devery Henderson, Greg Lewis, Billy McMullen, Jimmy Smith

By the way, remember the argument that Owens dropped more passes last year because of his hand injury? Not necessarily true. Game charters marked Owens with seven drops in 2005, and that was only in half a season. Muhammad led in 2005 with 13, followed by Bryant at 12, Burress at 11, and Justin McCareins and L.J. Smith at 10. Based simply on these two years, dropped passes doesn't actually seem like a very consistent stat. Only four receivers had at least seven drops in both 2005 and 2006: T.O., Chris Chambers, and -- Tom Brady won't be happy about this -- Donte' Stallworth and Benjamin Watson.

Posted by: Aaron Schatz on 22 Aug 2007

17 comments, Last at 22 Aug 2007, 10:00pm by Will Allen


by B (not verified) :: Wed, 08/22/2007 - 4:33pm

What about a drop percentage on third down? Maybe the reason TO had so few drops on third down is because Romo was looking elsewhere.

by andrew (not verified) :: Wed, 08/22/2007 - 4:41pm

Troy Williamson definitely deserves to be up there, especially given how few offensive chance the vikings had, and that he lost playing time (for his drops) as the year wore on, effectively doing most of his damage in the first half of the season (although he had one terrible one in Jackson's first game as a starter).

by David (not verified) :: Wed, 08/22/2007 - 4:43pm

Billy McMullen? I know it was the Eagles in 2005, but I'm amazed he even touched four passes in a year, much less dropped that many.

by mactbone (not verified) :: Wed, 08/22/2007 - 4:45pm

I knew Moose dropped a lot of passes in '05 but I wasn't sure how much of those were as killer as I remembered. Apparently they were.

by dbt (not verified) :: Wed, 08/22/2007 - 5:01pm

moose was probably too busy working on his sideline rants about kyle orton in his head to concentrate on catching the ball.

by A for anonymous (not verified) :: Wed, 08/22/2007 - 5:12pm

1) I agree. I am very interested in this article, but was slightly dissapointed because if TO was only targeted 4 times on third or fourth down, he's not exactly off the hook, especially when his name appears right next to Jerramy Stevens.

Also, I recall drops being charted based on the qb too, and since Terry Glenn was listed the year before, maybe the Bledsoe/Romo counts could have been addressed.

sorry for nitpicking maybe, and IMHO great analysis

by Will Allen (not verified) :: Wed, 08/22/2007 - 5:16pm

I don't know if we should think any better of a guy who dropped a fifteen yard pass on first down, rather than on third down. It seems that the down on which the incompetence was displayed is a random event.

I'm tellin' ya', in regards to the Vikings, Williamson wasn't the only receiver who killed their season with drops. I suspect that if the Vikings had just achieved the median number of drops, they may have won three or four more games. If Williamson had caught the ball, and thus stayed on the field with his speed, and kept opposing defensive coordinators from placing nine guys in the box, I can almost guarantee they would have won nine games.

by Harris (not verified) :: Wed, 08/22/2007 - 5:28pm

I'm most amazed that Billy McMullen saw enough passes on 3rd/4th downs to get that many drops. Or did he just drop everything that came his way?

by Rex (not verified) :: Wed, 08/22/2007 - 5:49pm

What about the perfectly thrown touchdown pass that TO dropped against the Redskins at Fedex field?

That ended up not only losing the game, but ended up having home-field advantage ramifications.

Can you do a sort based on catchable touchdown passes that were dropped? Or game situations, because a 1st quarter drop is clearly not as important as 4th quarter drops

by zlionsfan (not verified) :: Wed, 08/22/2007 - 6:05pm

I don't think they can get quite the statistic you might want: we chart the distance from the line of scrimmage when the pass was dropped, but that would only show you passes thrown into the end zone and dropped, not passes thrown over the defense and dropped.

by CA (not verified) :: Wed, 08/22/2007 - 6:25pm

Re: 7 I don’t know if we should think any better of a guy who dropped a fifteen yard pass on first down, rather than on third down. It seems that the down on which the incompetence was displayed is a random event.

Good point, Will. I think Aaron is feeding a bit into Simmons' clutch/choke obsession here. 3rd and 4th down drops generally are worse than 1st and 2nd down drops because they kill drives. That said, if anything, a disproportionate number of 3rd and 4th down drops, as Holt had, probably suggests that the player is likely to improve the following season, both because the down in which the drops occurred probably was random and because drops in general apparently aren't a consistent problem for most players from year to year. A large proportion of drops in "clutch" situations should be less a cause for criticism than a cause for hope. It seems to be mainly a product of bad luck that is unlikely to reoccur in the future. In that respect, this piece isn’t letting Owens off the hook at all: It indicates that he may (I stress, may) be one of the few players for whom drops are a consistent problem, and, unlike Holt, he does not have a particularly good chance of reducing the proportion of drops that occur on 3rd and 4th down by simple regression to the mean.

by Brian (not verified) :: Wed, 08/22/2007 - 6:28pm

Between the points made by #1 and #7 this stat is almost meaningless.

Drop per Target is probably the only meaningful way to slice it.

And #9, I disagree. Points in the first quarter count just as much as those in the fourth. Do wins in December mean more than wins in September?

by SoulardX (not verified) :: Wed, 08/22/2007 - 6:39pm

I suppose regression to the mean is good for the Rams recievers on this list--Holt, Bennett and McMichael.

by Pacifist Viking (not verified) :: Wed, 08/22/2007 - 6:50pm

Will Allen,

It is possible, though, that a drop would be considered a mental error rather than a physical error. In that case, it is possible a player might be more likely to make the error on a drive-saving or drive-breaking play rather than on first down.

Basically, I'm saying if Troy Williamson's drops last year were caused by poor vision, then he'd be as likely to drop a pass on any down. If the drops were caused by mental lapses ("chokes," I suppose), then the down could have an impact (though I only say "could," and I'm not sure there's a good way to statistically explain the likelihood of a mental lapse, anyhow).

by Pacifist Viking (not verified) :: Wed, 08/22/2007 - 7:20pm

I would guess that leaguewide, catching percentage is pretty similar across 1st, 2nd, and 3rd down (though I could see rational explanations for a difference: defense expecting the pass on 3rd down, defense allowing short completions on 3rd down, etc.). Just curious: has FO found a similarity or difference in catch percentage by down?

by SOW (not verified) :: Wed, 08/22/2007 - 9:06pm

Re: 9 and others.
First quarter drops and fourth quarter drops have the same amount of importance. What they don't have is the same amount of urgency. In the case of the first quarter drop, you still cost the team points, but you have the rest of the game to make up for it. In the case of fourth quarter drop, you don't have as many chances to do it again because the clock has run down. In both cases, the team potentially loses points and field position, but the timing makes one mistake much more difficult to overcome.

by Will Allen (not verified) :: Wed, 08/22/2007 - 10:00pm

Pacifist, the chance of getting a big enough sample size to estabish the validity of the third down drop/choke is pretty slim. I suppose it is possible, but I have my doubts.