In this week's Varsity Numbers, Bill Connelly takes a page out of baseball's playbook and attempts to isolate power from efficiency.
03 Aug 2010
by Bill Barnwell
Last week, I finished up a series on improving catch rate by introducing a statistic which compares a player's actual catch rate to a league-average figure, given the pass distance, relative location, and game situation.
However, catching the ball is just one aspect of a receiver's job. Once they have the ball in their hands, receivers are expected to pick up yards after the catch (YAC). Much like catch rate, though, YAC can be dependent on a receiver's usage pattern or subject to the whims of a quarterback's accuracy. With that in mind, I'm going to analyze YAC in the same way I did catch rate; I'm going to identify what the "expected" YAC should be for a pass, measure how players actually performed versus those expectations, and answer some questions about YAC and the skills of certain players and teams after the catch.
First, let's provide an example of how YAC+ calculates its baselines and compares expected YAC to actual YAC. We're trying to measure a player's ability when he actually catches the ball, not how much value he's providing the team on a per-attempt basis, so unlike our plus-minus statistic, YAC only considers completed passes.
Our sample uses four years of data (2006-09) from the Game Charting Project here at FO. The NFL does include both "Air Yards" and YAC with each pass play nowadays, but it's occasionally inaccurate; usually, there will be one set of downs every three or so halves I chart where the yards in the air and yards after catch are flipped, or mis-aligned, or just totally off. We instruct our charters to correct these mistakes where they're noticeable. Also note that air yards are the yards a pass travels in the air from the line of scrimmage, so a pass thrown three yards behind the line of scrimmage has the same distance in the air as one thrown six yards behind the line. Finally, all passes thrown into the end zone are marked with a maximum distance of (line of scrimmage - distance to end zone).
We construct the baseline for expected YAC as an average of two different measures; as our example play, we'll use an eight-yard pass to the left side of the field on second-and-10.
First, we find the expected YAC for an eight-yard pass to the left side of the field, regardless of the down and distance. That's 3.0 yards. Then, we find the expected YAC for a pass thrown to the left side of the field that's two yards short of the sticks on second down. The average YAC on such a completion is 3.4 yards. The average of these two figures is 3.2 yards. In reality, the 38 such completions over the past four years have yielded an average of 3.5 yards after catch, with most coming thanks to a 49-yard romp from Greg Jennings on a pass from Brett Favre in 2007. We calculate that Jennings produced 45.8 YAC+ on the play. Only one other receiver hit double digits on such a pass: Antonio Chatman.
For situations where the sample size is too small, we fill in expectation data from neighboring situations. Although the historical expected YAC for passes thrown 44 yards up the middle and 45 yards up the middle might be very different based on a small sample, I'm willing to guess that the historical expected YAC for all passes between 42-47 yards is a smoother, more accurate sample for the "true" expected YAC of those passes.
The tables below offer up some basic insight into how expected YAC changes through each of our modes of analysis. Figures that have an asterisk next to them mean that the number in question involves some estimating from neighboring cells.
|Table 1: Expected YAC Per Completion|
|Table 2: Expected YAC To Left Side By Down, Distance|
|Relative Yards||1st Down||2nd Down||3rd Down|
This is going to be more subtle than our work on catch rate, and the differences between YAC and YAC+ are going to be less pronounced than catch rate and plus-minus. In general, it's going to be hard for a player that picks up 40 yards after a catch to look bad in YAC+. There are some interesting findings, though, that we'll cover. Unless noted, all stats and rankings below are for players with a minimum of 30 targets for the year(s) in question.
One thing YAC+ does do is attempt to allow for comparisons between YAC gained by running backs and YAC gained by wide receivers. Last year, the top ten spots on the YAC charts were all occupied by running backs. On the YAC+ charts, four of the ten players are wideouts. It makes sense when we think about the players in question; was LeSean McCoy really more effective after the catch than Miles Austin? Table 3 compares our raw YAC leaders to YAC leaders.
|Table 3: Comparing YAC Top Tens|
|Top Ten, Raw YAC||YAC per Catch||Top Ten, YAC+||YAC+ per Catch|
|Darren Sproles||11.6||Darren Sproles||4.5|
|Chris Johnson||10.8||Hakeem Nicks||3.6|
|Julius Jones||9.9||Chris Johnson||3.3|
|Adrian Peterson||9.9||Adrian Peterson||3.2|
|Pierre Thomas||9.9||Patrick Crayton||3.0|
|Steve Slaton||9.9||Miles Austin||2.9|
|Justin Forsett||9.9||Kevin Smith||2.8|
|LeSean McCoy||9.9||Steve Slaton||2.6|
|Steve Slaton||9.3||Greg Jennings||2.4|
|Ray Rice||9.3||Justin Forsett||2.3|
It's not a surprise to see Sproles at the top of the list. He wouldn't have qualified for the 2008 list, as he only had 29 catches, but at 5.1 YAC+, he also would have led qualifying players in that category a season ago. Just behind him was Devery Henderson, at 4.1 YAC+.
Working our way to the other side of the rankings, only five of the 26 qualifying backs in 2009 put up negative YAC; perhaps I need to put an adjustment for the position of the player in question. Among those five backs -- and ahead of just Jamaal Charles -- is one Reggie Bush, whose 6.1 raw YAC per catch belies a YAC+ of -0.5 yards per catch. That's nothing new for Bush -- he had 1.2 YAC+ per catch as a rookie, -0.9 as a sophomore, and 1.2 YAC+ again in his third-year campaign. He is yet to have a season where he exhibits anything resembling significant production after the catch as a receiver.
The ten worst YAC per catch figures in 2010 all belonged to wide receivers; the YAC+ list introduces a couple of different wideouts, but only one player at a different position: tight end, where Donald Lee of the Packers wasn't contributing very much after he got the ball in his mitts.
|Table 4: Comparing YAC Bottom Tens|
|Bottom Ten, Raw YAC||YAC per Catch||Bottom Ten, YAC+||YAC+ per Catch|
|Malcom Floyd||2.5||Andre Caldwell||-1.6|
|Michael Jenkins||2.4||Bobby Wade||-1.7|
|Mark Clayton||2.4||Malcom Floyd||-1.8|
|Torry Holt||2.3||Steve Smith (NYG)||-1.8|
|Steve Smith (NYG)||2.3||Michael Jenkins||-1.9|
|Lee Evans||2.2||Donald Lee||-1.9|
|Kevin Walter||2.2||Greg Camarillo||-1.9|
|Greg Camarillo||2.0||Lee Evans||-2.0|
|Nate Washington||1.9||Nate Washington||-2.0|
|Ted Ginn||1.2||Ted Ginn||-2.8|
By either metric, Ted Ginn was the worst receiver after the catch in football last year, which is downright remarkable for a player who was supposed to be a threat downfield and with the ball in his hands. Ginn was also below-average in 2008 (YAC+ of -1.3) and just slightly above-average in 2007 (0.2). He remains the focal point of what is shaping up to be the worst draft in NFL history, the 2007 Dolphins draft run by Cam Cameron and Randy Mueller.
Also note the presence of Steve Smith on that list. Remember that Smith had the most catches above expectation of any player in the league in 2009, and the highest total of the past four years; simultaneously, he left more YAC on the field -- a cumulative total of -194.9 YAC -- than any player in the past four seasons. What a sublimely weird season, and one that seems unlikely to recur.
In Part II, I'll take a look at some more strange players and myths about YAC, and we'll examine whether quarterbacks have any control over the YAC gained by their receivers.
47 comments, Last at 13 Aug 2010, 8:21pm by Tim Wilson