Statistical In-Ferentz, Week 11: McBananas

Yesterday the ten finalists were announced for the Biletnikoff Award, and Marvin McNutt was, sadly, not on the list. The trophy, given annually to the "nation's most outstanding college football receiver" is awarded by the Tallahassee Quarterback Club, oddly enough, and the fine folks there have decided on the following players:

• Keenan Allen, Cal
• Justin Blackmon, Oklahoma State
• Ryan Broyles, Oklahoma
• Patrick Edwards, Houston
• Michael Floyd, Notre Dame
• A.J. Jenkins, Illinois
• Sammy Watkins, Clemson
• Jordan White, Western Michigan
• Robert Woods, USC
• Kendall Wright, Baylor

But how else are we to compare receivers except by counting stats like receptions and yards?  Without more detailed data, like targets, drops, and yards after catch, it is difficult, but one simple alternative is to calculate each player's stats as a percentage (of team catches/yards/TDs) rather than as a count. This way we can get a picture of how important a player is to his team's passing offense. It's not perfect -- do we really want to know the most important receiver on Air Force? -- but as a means of contextualizing the top candidates in their offenses, it is revealing: 

Now McNutt's already good stats start to look even better: he is second in terms of percentage of team receiving yards, fifth in terms of percentage of team receptions, and second in terms of team receiving touchdowns. Also note that Houston's Patrick Edwards suddenly drops off the radar in a couple of categories, simply because Houston passes so damn much (his yards per catch are impressive, though). Even Justin Blackmon starts to look a little less outstanding when you factor in how pass-heavy his team is.

An alternative to using percentages would be to standardize the numbers somehow. If we look at each player in terms of his team's passing attempts, there are wide gaps: Justin Blackmon's Oklahoma State team has attempted 459 passes, while A.J. Jenkins' Illinois team has attempted just 271. But we can adjust for this discrepancy by dividing yards, receptions and touchdowns by team pass attempts, then multiplying the resulting per play averages by some standard number of team pass plays. I chose 382, because that was the average number for the Biletnikoff list:

The standardized numbers tell an interesting story: now Marvin McNutt jumps up to second in yards, fifth in receptions and first in touchdowns. And the jump for A.J. Jenkins is even more dramatic. There are some important caveats here, though. First, I didn't compare all college football players, so who knows if there is some other undeserving low-usage wide receiver out there. Also, every one of these 11 players is on a team that passes at least somewhat. If you tried to standardize the receptions of all the receivers on, say, Ohio State or Navy, you would get some nonsensical results.

A more serious caveat involves something known as the usage/efficiency trade-off. This is a well-known concept from the world of basketball, where a player typically sacrifices some efficiency (field goal percentage) as his usage increases (i.e. as he takes more shots). This is why some statheads are not too impressed with gunners like Carmelo Anthony or Allen Iverson: they score a lot, but their efficiency declines as their shot total goes up. Something similar probably applies to football. Wide receivers are much more dependent on their quarterbacks and play-callers for opportunities than shooting guards, but it's reasonable to assume that defenses adjust somehow when offenses decide to target the same player or same type of player (e.g. receiver) over and over again, whether by sending a double team or playing more nickel and deep zone defenses. If that's the case, then it is not a simple matter of saying that Marvin McNutt's or A.J. Jenkins' receptions would scale up if Iowa or Illinois ran more passing plays relative to run plays. Instead we would expect that opposing defenses would adjust somehow to take away this new emphasis, with the effect that McNutt's or Jenkins' receptions, yards and touchdowns would still increase, but at a slower rate. So maybe A.J. Jenkins wouldn't have 1600 receiving yards, and maybe Marvin McNutt would have only 11 touchdowns instead of 13. Likewise, the players on faster-paced, passing heavy teams might benefit from a boost in efficiency if they played in offenses that ran the ball more. It's hard to know. But as a starting point for thinking of these players not just in terms of pure counting stat production but production within a system of opportunity, then the standardized figures are worth at least considering.

There are still other concerns to address, however, like broader issues of overall team quality: does a receiver have a good quarterback throwing him the ball, a good offensive line protecting that quarterback, a good running game/secondary receivers to keep the defense honest (or, alternately, such a bad running game/secondary receivers that passing to the star is the only option), and a good (or very bad) defense to get the ball back to the offense as quickly as possible? And there is the issue of the quality of the opposition: is a team playing a pillow-soft schedule, like Houston, or a truly hard schedule, like Oklahoma State, Baylor or Oklahoma (this criterion does not do much for Marvin's case, by the way). Maybe the lesson is that the position of wide receiver is a fundamentally contingent one: whether a wide receiver catches a lot of receptions depends to a great extent on factors beyond his control: system, teammates, pace, opposition, game situations. In which case the only fair way to determine the award, it seems obvious, is by number of boss one-handed catches. Check and mate, McNutt.

But if we must rely on statistics to some extent, then factoring in pace and style of offense would seem to be a necessity, and if we do consider those factors, then a Big 10 wide will be a top candidate to win the Biletnikoff award this year. Unfortunately, that wide receiver will not be Marvin McNutt, and not just because Marvin didn't make the final ten. It will be A.J. Jenkins. Considering the limited opportunities Jenkins had to catch the ball and the poor overall quality of the Illinois offense, his achievements, which were already impressive, begin to look fantastic. 7.6 receptions per game in an offense that only averages 17 completions per game? 113 yards per game for a team that averages 384 yards a game in total offense? Yeah, that's pretty good. Only his touchdown totals look a little suspect, but that might have something to do with the fact that Illinois is 81st in the country in red zone scoring attempts per game. And when you add in the requisite RZF* adjustment, his numbers look even better.

* Ron Zook factor, currently a 1.5 multiplier for yards, receptions and touchdowns.

But as far as just the top 10 goes, however, it's hard to see how Marvin McNutt doesn't make the list. McNutt's numbers are similar in absolute terms to those of Keenan Allen, Michael Floyd and Sammy Watkins, and as a percentage of team production are far better. McNutt's touchdown totals and yard per catch average are top-notch as well. The only reason I can see for McNutt's omission, at least statistically, is his relatively low reception and yardage totals, both of which are related to Iowa's slow pace and predilection to run the ball. That and the whole slow-Iowa-slow-Big-Ten-slow stereotype. But come on, Biletnikoff Award voters. Don't buy into superficial, gaudy, misleading stats. Despite what ESPN might tell you, numbers frequently lie, and two counting statistics in particular -- receptions and receiving yards -- paint a misleading picture of wide receiver production this year.