The Overrated Sack

Khaled Elsayed takes aim at the inflated value of sacks, and offers a few insights into how misleading the raw numbers can be when used to judge pass rushers.

| 4 years ago
overratedsacks-FEATURE2

The Overrated Sack


Every year I make it my mission in life to come up with new and creative ways to show the sack stat is the most abused one out there.

Sure, for some people it can be a reward for consistent play, but many others benefit from people caring too much about a small sample size, or, in some cases, players are punished because people don’t pay enough attention to what they actually do on the field.

Sample Size

Here’s what really gets me about the sack stat. You’re judging a player by what he does on a small percentage of his plays, and using that small percentage to assume a whole lot about the rest of his game. Now full disclaimer here, PFF doesn’t credit ‘half sacks’. No. we count a half sack as a full sack for an individual, so we’re a little bit kinder to pass rushers in this regard. But even so, have a look at the table below that represents how many plays on which a pass rusher records a sack.

 

Sacks by Total Defensive Snaps

PlayerTeamSacksDefensive SnapsSack % by SnapsPass RushesSack % by Pass Rush
J.J. WattHOU218932.35%5743.66%
Aldon SmithSF209522.10%5023.98%
Von MillerDEN199152.08%4374.35%
Cameron WakeMIA178771.94%5313.20%
Geno AtkinsCIN167592.11%4943.24%
Clay MatthewsGB146862.04%3563.93%
Charles JohnsonCAR147971.76%4922.85%
Julius PeppersCHI137581.72%4872.67%
Chris ClemonsSEA118371.31%4892.25%

 

J.J. Watt had an incredible year. In fact we’d go as far to say it was a year for the ages, and it earned him our inaugural Dwight Stephenson Award for the best player in football. But it’s a shame he needed to lead the league in sacks to build his case, because his year was about so much more.

Indeed, the table shows Watt picked up a sack on just 2.35% of his total defensive snaps, and that number doesn’t include plays nullified by penalty where stats aren’t recorded. Fortunately, by leading the league he got the praise he was due, but what about a guy like Aldon Smith, who finished second in the voting for defensive player of the year. Not one of the better run defenders in the league, he essentially earned his votes on the back of his play on 2.1% of his defensive snaps.

Is that fair?

People will say he played better than Von Miller because his sack number was higher. Yet they discount that Miller had a higher sack percentage and 12 more tackles for a loss in run defense or in coverage than Smith.

Time to Sack

People also fail to realize that not all sacks are created equal. The average time to sack in the NFL is 3.8 seconds, although this owes something to 305 instances of a sack taking five seconds or more. In our grading system we reward sacks that come under three seconds (and without a QB being forced into the lap of a defender) with a heavier positive grade.

Indeed, we don’t really know of many passing plays that require the QB to hold onto the ball for more than four seconds. So, you’re either looking at the passer not getting rid of the ball quick enough, or the coverage unit doing such a good job that he can’t. Either way it’s a sack that isn’t entirely down to the pass rusher, though he gets all the credit.

Let’s look at the table to see what it shows us.

 

Time from Snap to Sack

PlayerTeamUnder 3 Seconds3.1 to 4.1 seconds4 or more seconds
Von Miller, LBDEN973
J.J. Watt, DEHOU6114
Aldon Smith, LBSF687
Geno Atkins, DTCIN647
Julius Peppers, DECHI473
Cameron Wake, DEMIA395
Charles Johnson, DECAR347
Jared Allen, DEMIN345
Chris Clemons, DESEA335
Clay Matthews, LBGB257

 

Now, I don’t mean to continually promote Von Miller, but do you not think him getting nine sacks in less than three seconds is slightly more impressive than those who picked up theirs after four seconds?

Nature of the Sacks

However, even taking sample size and the speed of a sack in to account doesn’t make things crystal clear. You see, some sacks are earned and some are handed out. What about if a pass rusher actually had to beat a man to earn his sack? Sometimes a guy is left unblocked, sometimes the pressure from elsewhere pushes a quarterback into a player’s lap, and sometimes he just chases down a quarterback in pursuit who should get rid of the ball.

 

Adjusted Sack Count

PlayerTeamSacksUnblockedAdjusted Sacks
J.J. Watt, DEHOU21318
Aldon Smith, LBSF20515
Von Miller, LBDEN19514
Cameron Wake, DEMIA17314
Geno Atkins, DTCIN16313
Clay Matthews, LBGB14212
Julius Peppers, DECHI13112
Charles Johnson, DECAR1459
Chris Clemons, DESEA1129
Jared Allen, DEMIN1055

 

Look at these numbers closely. All of a sudden that magical number Aldon Smith produced doesn’t look quite so healthy, while the total Von Miller managed looks a little inflated. Meanwhile, poor old Julius Peppers got just the one freebie. That doesn’t make him a better pass rusher, but it should help you understand the sack stat is one that shouldn’t earn a great deal of your trust.

The Alternatives

We understand the need to quantify things with numbers. We do it ourselves. On the PFF Premium package you can go back to 2008 and check out numbers that don’t worship the small sample size produced by looking at sack numbers alone. We’ve got hits and hurries, and what’s more we’ve got the all important context — pass rushing snaps. Going beyond this we have our grades that take into account how quickly pressure happens and if a player had to do anything to get it.

The information is there. You don’t need to be a slave to the sack stat anymore and can instead appreciate players for more than what they do on 2% of plays.

 

Follow Khaled on Twitter: @PFF_Khaled


  • TundraGator

    Any place we can access stats like this for other players? Just curious.

    • MIRACLE

      Premium section mate

      • Tplum

        the stats are not all shown in premium stats. 

        • LightsOut85

          Hopefully they will be soon enough. Having them not available even when you’re willing to pay is such a tease.

      • TundraGator

        Premium section doesn’t have any of this info about times

  • Jeff Roy

    Khaled, is there a weighted scale that combines sacks, pressures and hits into an overall pass rushing grade?

    • Mauha Deeb

      Premium Stats have that.

      • Jeff Roy

        The Rush category for DE is compiled using their zero-based rating scale based on performance in individual plays. This is the only detail they provide on the “How We Grade” page.

        I guess the question I’m asking is what role does that have in determining the grade on a specific play. Obviously a sack is >  a hit > a pressure. I wonder how much?

        In this article, Monson mentioned how remarkable Watt’s stats are given that he is a interior lineman. Is a grade weighted by position, or should it be?

        • Mauha Deeb

          I don’t believe the grades themselves are weighted, but the general assumption would be that it is more remarkable when a DT(Atkins) or 3-4 DE(Watts) gets a +50 Pass rush grade compared to a pass rushing OLB(Miller).

          That is not a knock in anyway on Miller, BTW.

          • Jeff Roy

            I concur on the “weighting,” and would not advocate it. As NFL fans, I think we are reaching the limits of how much the game can and should be quantified.

            But when someone like Watt far exceeds the expectations of his position, you search for some means to measure it. PFF comes pretty close, as does Football Outsiders.

        • Arif Hasan

          Yes, they do have a section on weighting the impact on sacks vs. hits vs. hurries – it’s in their Pass Rusher Productivity Signature Stat section.

          On offense, they grade pass blocking with a Pass Blocking Efficiency score that weights sacks, hits and hurries differently and produces a number from 0-100 that determines their efficiency (they cluster around 95. I assume 0 theoretically means they allow a sack on every offensive snap).

          • Jeff Roy

            Thanks for making me to look a little deeper into the issue. Based on the raw numbers for 3-4 DEs, the actual weighting is not apparent.

            The difference in detail between that position and 4-3 DEs is striking. There are four 4-3 pass rushers with higher PRPs than Watt, which is forcing me to reconsider weighting for position.

            J.J.’s run responsibilities are greater due to the 3-4 alignment, which is reflected in the Run Stop % figures for each formation. Further proof that Watt had a season for the ages.

  • LightsOut85

    Is middle column in the time table supposed to be 3.1 to 3.9? (not 4.1), as the last column is “4 or more” (that would mean there’s overlap)

  • http://twitter.com/mr_gibson1 LaMar Gibson

    Are there stats on team defenses that excel at “coverage” sacks? You could probably infer from other statistics but it would be cool to see which secondaries really boost the pass rushers’ numbers.

  • Wmdegrange

    couldn’t we infer that a player that gets 21 sacks probably also is very good at getting pressure, getting to the qb quickly, and beating blocks? After all this is the NFL they are not largely giving away sacks.

    • Jcorye1

      Eh, you could also argue he got incredibly lucky multiple times.

  • http://www.facebook.com/people/Benjamin-F-Barnes-Sr/717949335 Benjamin F. Barnes Sr.

    What about double teams and TEs or RBs blocking. If Adkins is coming from inside and getting all double teams that is impressive.

  • Izach

    interesting to see the time to sack numbers, it actually works to show clay matthews is a real try hard guy and a never say never pass rusher  since most of his sacks are second effeort or “coverage” sacks

  • pbskids4000

    I’m pretty sure Von Miller came in second in voting for the DPOY.

    • broncosfan18

      He did

  • http://www.facebook.com/joseph.e.obrien Joseph E. O’Brien

    How does Lawrence Taylor and Reggie White and Bruce Smith compare to these players in terms of Sack % by Pass Rush. Has anyone looked back at some of the great pass rushers of the last 20 – 30 years and compared them to today’s players. I always thought that if LT played in today’s game he would hit 25 sacks easily…

  • D Crook

    JJ Abrams SWATTED DOWN a few more passes than any other defensive player in the NFL…!  Not as devastating as a sack, but it sure wrecks the QB’s momentum…!  It might be interesting to survey QB’s & O-linemen to find out who THEY think is toughest & most disruptive to play against…! 

    • JonFrum

      A blocked pass is in incomplete pass. A sack is a loss of yardage.

  • omar

    hey where’s demars ware? i’m surprised he didn’t appear, i know he had a down year and just 11.5 sacks but he seriously he didn’t have more than two sacks under three seconds? 

  • Mark7425

    While the sack is not a very relevant stat, I think the rate of hurries:sacks or hits:sacks, etc. would be very interesting. Like a stat that shows how many hits for every sack or how may hurries for every sack.

  • Arthur Jackson

    It would seem to me as a non advanced stats metrics guy that some some of the longer sacks could be as a result of chasing the QB down, especially away from you with the added speed of a rusher flustering a QB into not being able to unload it to and open receiver, while not throwing it away because of the open receiver which would mean that it is not a coverage sack. Looking at unblocked sacks isn’t it also possible that the less unblocked sacks a players has would also correspond to being chippped and double more often also? I guess my feeling is that these are good numbers presented here and give a more complete picture, in some aspects, but also raise more questions.