A week after big upsets of Stanford and Ohio State, the USC and Virginia Tech themselves fell to less-talented opponents. Georgia also fell to South Carolina after pummeling Clemson in the opener.
11 Dec 2011
by Aaron Schatz
Earlier this week I wrote an article for ESPN Insider about Jordy Nelson's great season. In that piece I debuted in-season receiving plus/minus numbers for the first time, as well as YAC+ numbers. As promised, here are some more of those numbers for FO readers to check out.
Bill Barnwell introduced receiving plus/minus back in 2009. The idea was to adjust catch rate for the fact that different receivers run different types of routes, and deeper balls aren't going to be caught as often as short passes. Each catch is compared to a baseline of how often balls are caught based on the length of the pass and whether the pass is to the left, middle, or right. Now, the original plus/minus used game charting data to remove uncatchable passes: passes marked "Thrown Away," "Tipped at Line," or "Hit in Motion." If we want to do these numbers in-season, we run into the problem where our game charting numbers are always a couple of weeks behind the standard play-by-play numbers. Therefore, I created new baselines which are based on all passes without adjusting for game charting data. These numbers won't be as accurate, but they do give us a basic plus/minus to use in season.
There's a second problem with doing plus/minus in season, and that's the fact that there are often mistakes by official scorers when it comes to PYD/YAC. These are "unofficial" stats and therefore there isn't the same pressure from the league to get them right. When the official scorers are trying to track 20 things on each play, there are going to be some mistakes. Some official scorers make more mistakes than others. Usually, we correct these mistakes in the course of game charting, and then report those fixes to the league and get the data changed in league data. However, I have not yet made changes in the FO data unless we've officially corrected that data with the league, so there are some mistakes in here. The league as a whole comes out at -189.6; I have a feeling that this comes from those plays where the official scorer doesn't enter the PYD/YAC correctly, and thus longer passes get listed as 0 PYD. There's also a general problem with incomplete passes marked as "0 PYD," because passes thrown away or tipped at the line are often just marked as 0 by the official scorers. But we'll live with those issues for now, knowing that we'll gradually be able to fix the data over the next few weeks, and then get the game charting data finished and do more accurate plus/minus numbers.
Obviously this isn't a perfect measure of receiving. Besides the problems listed above having to do with the limitations of measuring this in-season, we have a couple of other issues. We're not measuring the specific routes here; I'm sure there's a difference between how often a 15-yard dig is caught and how often a 15-yard curl is caught. We don't have that data for each pass. In addition, obviously the abilities of the quarterback play a big role in the plus/minus numbers, just as they do in catch rate. Green Bay and New England receivers are generally high; Cleveland and Jacksonville receivers are generally low. We could adjust these numbers for a team baseline, but we won't do that this week; besides that presents its own problems -- for example, why should Jordy Nelson's numbers be discounted because Greg Jennings and Donald Driver are also excellent recievers.
Let's take a look at the best and worst players so far this year according to in-season receiving plus/minus.
|Top 20 In-Season Receiving Plus/Minus, 2011 Weeks 1-13|
|Bottom 20 In-Season Receiving Plus/Minus, 2011 Weeks 1-13|
I'll save the in-season YAC+ numbers to run later this week in a separate post.
14 comments, Last at 17 Dec 2011, 7:25pm by Mr. Guest to you