Much has been made recently on television (especially during the Cleveland Browns game as Braylon Edwards was also on the field) about Brandon Marshall's drops.  Does he really drop the ball more than he should?

As a disclaimer, let us first get out of the way the idea that a dropped pass is a subjective measure, and as such isn't a fully accurate statistic, like yardage or number of touchdowns which can clearly be shown to be an exact (or at least rounded to nearest yard) value.  The data I found on number of drops was provided by Stats, Inc and I could not find their definition of a drop anywhere, so technically the could be counting pretty much anything.  I think, however, most reasonable individuals would define a drop as something along the lines of "a pass thrown to a receiver which is not recepted (by the receiver in question) but is touched by both hands of the receiver."  If anyone CAN find Stats, Inc's definition, please do share!

First, a look at raw number of drops.  Brandon Marshall is third in the NFL in drops with 11, behind Dwayne Bowe (13) and Braylon Edwards (16!).  However, as many of us may know, Marshall also leads the NFL in "targets" (number of times thrown to, another subjective, albiet less so than drops, stat), with 146 targets, eight above Andre Johnson who has even played one more game.  So is it really fair to say, based on raw number of drops, that Marshall drops too many passes?  Nay I say.

I went looking for a "drop rate" stat, and the only one I stumbled upon was receptions per target.  This isn't quite the same though; many receivers are deep threats, and do not get as many catchable passes as a possession receiver or someone like Joseph Addai who is constantly receiving checkdowns from Manning.  We should be proud to note, however, that Eddie Royal ranked 12th and Tony Scheffler ranked 18th in this catch per target category among players with 3.125 targets per game in the AFC!

Anyways, to concoct the results below I took the 20 players with the most drops from the AFC and from the NFC, and divided their drops by their number of targets, to obtain a drop rate.  Note that you could also use these results to determine "uncatchable target rate", another interesting stat.

AFC

NFC

Conclusions

As you can see, discounting poor Madison "Butterfingers" Hedgecock who only has 12 targets anyways, Braylon Edwards (who anecdotally was dropping too many passes) and Marques Colston "lead" their respective conferences in drops per target.  Clearly our list also incorporates some of the best in the league, as I haven't heard anyone complaining about Steve Smith's, Santana Moss' or Randy Moss' hands.  The list includes all pass catchers, thus you see some RBs and TEs in there.  I don't think this in any way taints our results, because RBs are usually getting short passes and shouldn't drop any more (in my opinioin) than a WR.  The TEs that appeared on the list are either great pass-catching TEs or talented youngsters and thus should (also in my opinion) not have a worse drop rate than average receivers either.  At any rate, it is doubtful that conducting this study with only WRs would yield any significantly different results.

For these 40 samples, the expected value for drop rate is 0.0924; on average we would see 9.24% of passes thrown at these receivers dropped.  The variance is 0.0048 and thus standard deviation (square root of variance) is 0.0696, meaning we would expect that over two thirds of the NFL would have a drop rate from about 0.03 to 0.16.

Furthermore, the Central Limit Theorem allows us to estimate the error in these calculations based on our sample size (40).  The average of the samples is 0.0924, and Marshall's drop rate of 0.0753 is below this average.  How much does the average of our sample predict the average of the league?  Well, the probability that the actual league average drop rate is greater than Marshall's drop rate of 0.0753 is 0.931.  Therefore, we can say with about 93% certainty that Marshall does not drop more passes per target than the average NFL player.