When analyzing specific defensive players, the problem with statistics like tackles and sacks is that they’re not context-dependent. In other words, a defender gets the same credit for a tackle two yards downfield whether he’s making that tackle on first-and-ten or third and two. It’s the same on offense – should a running back get the same credit for a four-yard run on second-and-twelve that he does on third-and-three?
At Football Outsiders, it’s our goal to bring as much in-game context to those base stats as possible. To that end, over the last few years, we have been assembling official play-by-play data for the things that it can reveal. In addition, between out staff and an army of volunteers, we’ve been charting every NFL game for three seasons. This allows us to get a much better handle on things like offensive formations, quarterback actions, the involvement of specific defenders, incomplete passes, and blown blocks. The goal is to bring a new level of statistical analysis to a game that didn’t start recording quarterback sacks until 1982. We’re making inroads, and the defensive stats we’ve invented are the insightful byproducts.
For the purposes of this article, we’ll be using these FO stats:
Plays: The total number of defensive plays in which a player is involved, This includes tackles, pass deflections, interceptions, fumbles forced, and fumble recoveries.
Percentage of Team Plays: The percentage of team plays involving this defender.
Defeats: The total number of plays which prevent an offense from gains first down yardage on third or fourth down, stop the offense behind the line of scrimmage, or result in a fumble (regardless of which team recovers) r interception.
Stop Rate -- The percentage of all plays that are stops; i.e., the number of plays in which a defense prevents a success, based on 45 percent of needed yards on first down, 60 percent on second down, and 100 percent on third down. We’ll show this for overall and for Run Stops.
Average Yards – The average number of yards gained by the offense when this player is credited with making the play. We’ll do this for overall and against the run.
Linebackers have a few specific stats, which are proprietary and created by our game charting data. There are Targets (the number of pass plays in which the player was listed in coverage), Success Rate (the percentage of plays in which passing successes were prevented – same numbers per down as Stop Rate), and Adjusted Yards per Pass (the average number of yards gained on plays on which this defender was in coverage, adjusted for the quality of the receiver covered). These stats can be found in Pro Football Prospectus 2008, our season preview book that was just released this week (Plug! Plug!), and we’re gong to focus more on run stops because we’re talking about inside guys.
Opponent adjustment is a big part of what we do, and the commanding principle behind our DVOA and DYAR stats. Obviously, it’s more impressive to shut Randy Moss down then, say, Arnaz Battle or Keary Colbert. Plugging a hole that Adrian Peterson’s running through is going to get you far more “ooohs” and “ahhhs” than blowing out Shaun Alexander, who’s been shutting himself down for two years.
So, that’s the preamble. Let’s get specific, on the request of Art Bietz and all who follow him, and talk about three Cleveland inside linebackers. We’re going to compare and contrast with other inside linebackers who play in 3-4 defenses, to get as clear a read as possible. Rankings will eliminate all linebackers who play primarily in 4-3 sets. This leaves us with 22 primary (full-time starting) inside linebackers.
Plays/Pct (NFL Leader: Ray Lewis, 132)
Andra Davis – 70 (Rank: 15, 8.3%)
D’Qwell Jackson – 105 (Rank: 4, 14.2%)
Leon Williams – 83 (Rank: 13, 9.8%)
Stop Rate (NFL leader: Bart Scott, 72%)
Andra Davis – 66% (Rank: 3)
D’Qwell Jackson – 58% (Rank: 17)
Leon Williams – 60% (Rank: 16)
Run Stop Rate (NFL Leader: Bart Scott, 82%)
Andra Davis – 76% (Rank: 3)
D’Qwell Jackson – 52% (Rank: 15)
Leon Williams – 51% (Rank: 16)
Average Yards per Play (NFL Leader: Tully Banta-Cain, 2.5)
Andra Davis – 3.4 (Rank: 8)
D’Qwell Jackson – 5.5 (Rank: 21)
Leon Williams – 5.2 (Rank: 18)
Average Yards per Run (NFL Leader: Bart Scott, 1.8)
Andra Davis – 2.1 (Rank: 3)
D’Qwell Jackson – 4.3 (Rank: 18)
Leon Williams – 3.1 (Rank: 11)
Defeats (NFL Leader: James Farrior, 26)
Andra Davis – 15 (Rank: 13)
D’Qwell Jackson – 14 (Rank: 12)
Leon Williams – 11 (Rank: 14)
Now, here’s the problem with defensive stats above and beyond even the ones you just read: Unlike in baseball, there’s no way to completely isolate the performance of a single linebacker form the performances of those around him. So, if Cleveland’s interior defensive line was a problem in 2007 (it was, judging by Phil Savage’s desire to spend eleventy billion dollars to upgrade it), that’s going to affect the number of plays in which Cleveland’s linebackers are overwhelmed by offensive linemen or other blockers. Conversely, in a 3-4 defense, if you have a Shawne Merriman or DeMarcus Ware (Kamerion Wimbley seems to be on the precipice of excellence) grabbing double teams as a matter of offensive survival, that opens things up for those behind and around him.
Clearly, Andra Davis is the best of the three linebackers detailed here from a run-stopping perspective, but remember – so many things go into these numbers if you want to know what they really mean. The hope in Cleveland is that, one year after completely redefining his offensive line (which I wrote about here), Savage will do the same on the other side of the ball. If that’s the case, all three of these players could see their totals looking better.
Check out Football Outsiders and Pro Football Prospectus 2008 for more on the Browns and the NFL, and feel free to shoot me a note at doug_at_footballoutsiders.com if you have any questions.