Frederick Breedon - Getty Images
Team defensive stats are usually limited to production numbers by opposing offenses. Splash Value is a measurement of what defenses do to offenses, with different weights assigned to multiple defensive statistics.
Editor's note: In further researching this topic, I realized the stats pulled for sacks, interceptions, forced fumbles and fumble recoveries were those against a team. For whatever reason, ESPN lists those under "defense," which is insanely stupid, but it also renders these numbers invalid. I apologize for the error, I'm working on a comparative review of this season and the last two seasons with the correct data. Hoping to have that up 10/15. -nc
I don't claim membership in the Society of Stats Geeks. I've added two and two and gotten five many times. As far as statistics go, we know they will often tell us what we want them to.
My issue has always been whether that makes the statistic any more or less true. As sabermetrician Bill James said, the point of statistics is to illustrate something we already know. They aren't meant to predict.
The gatekeepers of statistics are the media, by and large. As a writer, my sources of statistical compilation include the same sources millions of fans, Fantasy Football gurus, nerds and other writers use. ESPN, NFL, Pro Football Reference, Pro Football Focus, Advanced NFL stats, and on and on.
What I mean by "gatekeepers," is the stats we will accept as being the most important are the ones they're going to default to. If you open up the Total Team Defense stats link on ESPN.com, the results return sorted by total yards allowed in ascending order.
This seems silly, considering bye weeks greatly skew that number, as does the slew of prime time games the league plays (i.e. the Lions have played four games and the Titans have played six, as of 7 a.m. Oct. 14). The bigger point here is according to ESPN.com, whether intentional or otherwise, millions of followers of the league that click on ESPN's stat page are subtly told the "best" defense in the league is the one surrendering the least amount of total yards.
There are plenty of arguments to be made against that line of thinking, but in the end, total yards allowed is just as arbitrary as any other, and ESPN has to default in some way, so I'm not suggesting this is categorically misleading.
It's one-sided, though. What it shows is more based in the result of what the offense did. With the Thursday Night game already played, the total yards stat on ESPN.com shows the Dallas Cowboys first in the NFL, having allowed 1,110 yards. It's a fact they've allowed that total, but that tells us what offenses have done against it; it doesn't tell us what the Cowboys defense has done to opposing offenses.
There's a more detailed level here. Stats that measure attacking or "offensive" plays by the defense tell another story, and one equally important.
Those plays don't measure the results of an offense against a defense. They highlight the big play ability of a defense. These are often referred to as "splash plays."
Giving team defensive statistics based on opportunities and splash plays created (with a weight) provides insight into which defenses are forcing plays into their advantage.
"Splash plays" for the purpose of this statistic are sacks, interceptions, passes defensed, forced fumbles and fumble recoveries. Simply put, good things happen when a defense does these things, and when a defense does these things, good things happen.
This does not necessarily trump any other statistic, but it provides a deeper look into the amount of plays where a defense did something measurable, instead of measuring the defense's effort as a byproduct of what the offense did. A reasonably-minded person would agree that a sack has more value than a two-yard gain. Both are positive defensive statistics, but the sack is more indicative of something the defense did, as opposed to measuring the defense based on what the offense did.
Teams are throwing the ball more than they are running it, and the two most positive results of those plays for the defense are interceptions and sacks. Since interceptions result in change of possession, they are given the most weight - (INT*3).
Sacks can really hinder an offensive series, and the defense of those can oftentimes be the difference in a game, so they have an increased weight - (SACK*2).
Passes defensed doesn't carry with it the same game-changing characteristic as sacks and picks, but a defensive player is credited with a PD when he makes a play on the ball, which is really the root of the splash play - (PD*0.5).
Forced fumbles are a little bit trickier. They carry with them the potential of a turnover, but without the ensuing fumble recovery, it doesn't tell us directly the team made an impact on that play. Let's say the Steelers give up a 40-yard pass play to Bengals WR A.J. Green in Week 7 (a likely scenario), and after that, he is hit by Ryan Clark, fumbles the ball but recovers it.
Is that a splash play? Seems more like a splash opportunity that wasn't realized. Still, in the spirit of this statistic tracking a big play made by the defense, and the fact there's obviously a correlation between the amount of fumbles forced and the amount of fumbles recovered, it seems inclusion is appropriate, if not given a lesser weight.
We're giving forced fumbles an even weight (FF+1) and fumble recoveries a weight that, when combined with forced fumbles, would equal an interception - (FR*2). A fumble recovery is given a lesser weight as a stand-alone stat than an interception to account for unforced fumbles.
A good example of that is Rashard Mendenhall losing grip of the ball in the open field in Week 5 against Philadelphia. Had the Eagles recovered that, they'd get credited with a fumble recovery, but not a forced fumble. The absence of the forced fumble, but the presence of the recovery still gets two points, but not the full three an interception would because they didn't fully make an impact by their own doing.
This statistic is not measuring the results of those plays - as in touchdowns and yardage gained after turnovers is not factored into the equation. Although it certainly could be, the purpose is to evaluate what they are doing on defense, and what happens after a turnover is still technically a defensive statistic, it becomes an offensive play.
While this will likely be the controversial piece of this stat, I'll illustrate it this way. If Ryan Clark gets an interception, and during his run-back, Ike Taylor (sorry Ike, I had to) gets flagged for holding, the official will often say, "after change of possession," or "during the return." The Steelers would still keep the ball, but get 10 yards marched off from the spot of Taylor's foul.
And with that, we give you the Week 6 NFL Defensive Splash Value leaders, along with their ranks in points allowed and yards allowed per game, heading into the Sunday and Monday games of Week 6.
|TEAM||Splash Value||Splash Value per game||RK||Scoring Rank||Yardage Rank||SACK||PD||INT||TD||FF||REC||TD|
*Through 7 a.m. Oct. 14.
The numbers bear out fairly closely with their points against and yards allowed ranks in the league, with a few aberrations. Washington ranking 11th in Splash Value despite sitting at 28 in both scoring and yards allowed is probably the most noticeable. Their six interceptions and league-high 36 passes defensed are the reasons for that. Again, this stat isn't showing the results of what offenses did to a defense (which is what yards and scoring measures). This is what defenses are doing to offenses.
As far as the Steelers go, sitting at 22 in Splash Value per game "seems" right, considering their low number of interceptions (two). It also fits in with their overall ranking of 19 in yards allowed and scoring. We've often lamented over their lack of "big plays," and this statistic bears that out.
Speaking of "Bears," sitting at No. 1 by a wide margin, shows the value of interceptions - they lead the NFL with 13 picks.
I have my own critiques of this, and am currently running this formula on past seasons to see what comes out. I don't consider this to be a final product at all, and I reserve the right to tweak and modify. I'd love to hear your thoughts on this (cleaner formats will be available in the future), so leave a comment, or hit me up via email or Twitter.