Adjusted Sacks and IDP Regression

Jeff Ratcliffe digs a little deeper into the 2012 sack totals and introduces adjusted sacks as a way to gauge regression toward the mean.

| 3 years ago
yahoo_aldonsmithsackdance

Adjusted Sacks and IDP Regression


There’s been a lot of talk about sacks over the last few weeks. Aldon Smith of the 49ers wants the sack record, but said he wouldn’t “punch himself in the face” if he doesn’t get it. Panthers DE Greg Hardy said his personal goal is 50 sacks this season, but he had no comment on self-inflicted face punching.

Hardy’s lofty goal is more than double Michael Strahan’s single-season record of 22.5 sacks. While I admire his ambition, it’s obviously an impossible feat. But Strahan’s record is certainly within reach. J.J. Watt, Aldon Smith, and Von Miller all made serious runs at the record last season, and Jared Allen came up just a half-sack short in 2011. 

These inflated sack numbers have really intrigued me. However, as someone who’s constantly looking at defensive player statistics, I can tell you that the raw numbers rarely tell the whole story. I’ve talked about this problem before with QB pressure frequency, but this metric still doesn’t give the depth of analysis we need to truly understand sack production.

To get a sense of how pass rushers produce, I put last season’s sack numbers under a microscope. In my analysis, one question kept popping into my head: How would defensive ends and rush outside linebackers produce in a vacuum? In other words, some players recorded sacks more frequently than others. What ultimately was the average conversion rate, and how would last season’s sack totals change if we apply the same conversion rate to everyone’s sack totals?

In order to determine the answers to these questions, I first conducted a five-year study where I looked at the QB pressures (the sum of sacks, hits, and hurries) accumulated by defensive ends and rush outside linebackers. I then calculated the sack conversion rate by dividing the sack total by the QB pressure total.

Year Sacks Hits Hurries QB Pres Conv%
2012 909 1101 3346 5356 17.0%
2011 924 1001 3448 5373 17.2%
2010 834 972 3550 5356 15.6%
2009 871 1195 3209 5275 16.5%
2008 853 1104 3260 5217 16.4%

The above data gives a solid sense of the league-wide trends among the pass rushers. Generally speaking, we’ve seen an increase in sack conversion rates over the last five years. While our sample size isn’t massive, 2010 does seem to be a bit of an outlier. The 15.6 percent conversion rate is nearly a full percentage point lower than the five-year average of 16.5 percent.

Now, with this data we can get to work adjusting last season’s sack totals. Remember that the goal here is to see how every pass rusher would perform on an even playing field. We accomplish this parity by assuming that all players convert their QB pressures for sacks at the same rate. This means applying the five-year average of 16.5 percent to each player’s pressure total from last season. The following displays all defensive ends and rush outside linebackers who recorded at least 10 QB pressures last season:

PlayerPosSnapQBPSackAdj SackDif
Cameron WakeDE877861514.19-0.81
Von MillerOLB9178618.514.19-4.31
Chris LongDE8477611.512.541.04
Charles JohnsonDE7987612.512.540.04
J.J. WattDE8957620.512.54-7.96
Jared AllenDE1004741212.210.21
Derrick MorganDE879726.511.885.38
Michael BennettDE91472911.882.88
Aldon SmithOLB9567019.511.55-7.95
Ryan KerriganOLB1023698.511.3852.885
Greg HardyDE71960119.9-1.1
Elvis DumervilDE88060119.9-1.1
Chris ClemonsDE8405911.59.735-1.765
Kamerion WimbleyDE8725869.573.57
DeMarcus WareOLB8295811.59.57-1.93
Julius PeppersDE7595711.59.405-2.095
Brian RobisonDE822568.59.240.74
John AbrahamDE71056109.24-0.76
Mario WilliamsDE8895610.59.24-1.26
Jason Pierre-PaulDE844556.59.0752.575
Paul KrugerDE7505599.0750.075
Michael D. JohnsonDE8235511.59.075-2.425
Lamarr HoustonDE826544.58.914.41
Israel IdonijeDE688517.58.4150.915
Carlos DunlapDE5705068.252.25
Ahmad BrooksOLB933506.58.251.75
Clay MatthewsOLB68749138.085-4.915
Dwight FreeneyDE6644757.7552.755
Trent ColeDE6964637.594.59
Jason BabinDE618465.57.592.09
Tamba HaliOLB8944697.59-1.41
Osi UmenyioraDE6154567.4251.425
Cameron JordanDE9934587.425-0.575
Robert QuinnDE8014510.57.425-3.075
Justin HoustonOLB96744107.26-2.74
Chandler JonesDE6984367.0951.095
Everson GriffenDE5784387.095-0.905
Jason HatcherDE7234246.932.93
Rob NinkovichDE8624186.765-1.235
Ray McDonaldDE920402.56.64.1
Connor BarwinOLB9514036.63.6
Calais CampbellDE726406.56.60.1
Anthony SpencerOLB81340116.6-4.4
Shaun PhillipsOLB800399.56.435-3.065
Muhammad WilkersonDE8703756.1051.105
Jared OdrickDE8873756.1051.105
Will SmithDE9443766.1050.105
Jabaal SheardDE9283776.105-0.895
Kroy BiermannDE6583645.941.94
Frank AlexanderDE529352.55.7753.275
Desmond BryantDE6113535.7752.775
Jeremy MinceyDE9193535.7752.775
Cliff AvrilDE663359.55.775-3.725
Sam AchoOLB9603445.611.61
Brett KeiselDE830334.55.4450.945
Kendall ReyesDE503335.55.445-0.055
Corey LiugetDE6863375.445-1.555
Calvin PaceOLB9613235.282.28
Justin TuckDE6253245.281.28
Darnell DockettDE763311.55.1153.615
Justin SmithDE7833135.1152.115
Corey WoottonDE5573175.115-1.885
Philip WheelerOLB9773034.951.95
Stephen BowenDE7422914.7853.785
Kyle Vanden BoschDE623293.54.7851.285
Robert MathisOLB5942984.785-3.215
Arthur JonesDE504284.54.620.12
James HarrisonOLB7852864.62-1.38
Jerry HughesOLB5622744.4550.455
LaMarr WoodleyOLB6012644.290.29
Daniel Te'o-NesheimDE6972644.290.29
Robert GeathersDE6222534.1251.125
Derek WolfeDE8552564.125-1.875
Brooks ReedOLB564242.53.961.46
Erik WaldenOLB7322333.7950.795
Mathias KiwanukaOLB5102233.630.63
Courtney UpshawOLB709211.53.4651.965
Frostee RuckerDE5632143.465-0.535
Akeem AyersOLB8332163.465-2.535
Jarvis JenkinsDE5281702.8052.805
Rob JacksonOLB564174.52.805-1.695
Cory ReddingDE5531622.640.64
Matt ShaughnessyDE643153.52.475-1.025
Ziggy HoodDE7761432.31-0.69
Mike DeVitoDE6061211.980.98
Red BryantDE6031101.8151.815
Russell AllenOLB966110.51.8151.315
Tyson JacksonDE580731.155-1.845
Ryan PickettDE546600.990.99

 

It’s immediately apparent that the sack totals at the top are much lower than what we actually saw. Cameron Wake and Von Miller top the list with an adjusted sack total of 14.2, which is significantly less than Watt’s actual total of 20.5 sacks. This means we obviously have some overachievers in the bunch. Here are the most notable overachievers who saw significant snaps last season:

Player Pos QBP Sack Adj Sack Dif
J.J. Watt DE 76 20.5 12.5 -8.0
Aldon Smith OLB 70 19.5 11.6 -7.9
Clay Matthews OLB 49 13 8.1 -4.9
Anthony Spencer OLB 40 11 6.6 -4.4
Von Miller OLB 86 18.5 14.2 -4.3
Cliff Avril DE 35 9.5 5.8 -3.7
Robert Mathis OLB 29 8 4.8 -3.2
Robert Quinn DE 45 10.5 7.4 -3.1
Shaun Phillips OLB 39 9.5 6.4 -3.1
Justin Houston OLB 44 10 7.3 -2.7
Michael D. Johnson DE 55 11.5 9.1 -2.4
Julius Peppers DE 57 11.5 9.4 -2.1
DeMarcus Ware OLB 58 11.5 9.6 -1.9
Derek Wolfe DE 25 6 4.1 -1.9
Chris Clemons DE 59 11.5 9.7 -1.8

Watt tops the list, converting nearly 27 percent of his pressures for sacks. He’s followed closely by Smith, who recorded a sack on almost 28 percent of his pressures. The rest of the adjustments aren’t as significant as Smith’s and Watt’s, though a number of these players will garner strong consideration in IDP drafts this season.

So what about the other side of the coin? Let’s take a look at the notable underachievers:

Player Pos QBP Sack Adj Sack Dif
Derrick Morgan DE 72 6.5 11.9 5.4
Trent Cole DE 46 3 7.6 4.6
Lamarr Houston DE 54 4.5 8.9 4.4
Connor Barwin OLB 40 3 6.6 3.6
Kamerion Wimbley DE 58 6 9.6 3.6
Ryan Kerrigan OLB 69 8.5 11.4 2.9
Michael Bennett DE 72 9 11.9 2.9
Desmond Bryant DE 35 3 5.8 2.8
Jeremy Mincey DE 35 3 5.8 2.8
Dwight Freeney DE 47 5 7.8 2.8
Jason Pierre-Paul DE 55 6.5 9.1 2.6
Calvin Pace OLB 32 3 5.3 2.3
Carlos Dunlap DE 50 6 8.3 2.3
Justin Smith DE 31 3 5.1 2.1
Jason Babin DE 46 5.5 7.6 2.1

Derrick Morgan tops the list, and his adjusted sack total of 11.9 should give you a strong sense of why many IDP analysts have pegged Morgan for a breakout in 2013. Morgan’s teammate Kamerion Wimbley also underperformed recording a sack on just 10.3 percent of his pressures. Likewise, Lamarr Houston’s low conversion rate of just 8.3 percent should regress upward toward the mean. He’s a sneaky value in fantasy drafts this season.

Further down the list, Ryan Kerrigan had an impressive 69 pressures. He’s an extremely talented pass rusher who is still flying under the radar. Snatch him up if you’re in leagues with big-play scoring systems. We also see a few notable underachievers on the list including Jason Pierre-Paul, Carlos Dunlap, and Jason Babin. But what do we make of their underachieving?

 

Can adjusted sacks be predictive of regression toward the mean?

So far, we’ve taken a look at last season’s adjusted sack totals, but what about the totals from previous seasons? Does the adjusted sack total give us a better sense of future performance, or are some players just better at converting sacks than others?

The answer to this question is certainly complex, and I always recommend closely watching game footage to get a sense of how effective a particular player is as a pass rusher. There’s no doubt that some are going to be better than others. At the same time, the numbers can be very insightful.

I’m not going to do a full regression analysis here, though that may be the topic of a future piece. Instead, let’s focus on the sack leaders from previous seasons. We’ll adjust their sacks and then compare that total to what they did in the following season:

Player Pos QBP Sack Adj Sack 2012
Jared Allen DE 66 22 10.9 12
DeMarcus Ware OLB 72 19.5 11.9 11.5
Jason Babin DE 67 18 11.1 5.5
Jason Pierre-Paul DE 55 16.5 9.1 6.5
Aldon Smith OLB 64 14 10.6 19.5

Among the 2011 leaders, Smith was the only player who did not drop off last season. The conversion rate from the other four players regressed toward the mean, with Babin and Pierre-Paul being the most significant examples. Let’s also take a quick look at the sack leaders from 2010 and what they did the following season:

Player Pos QBP Sack Adj Sack 2011
DeMarcus Ware OLB 78 15.5 12.9 19.5
Tamba Hali OLB 97 14.5 16.0 12
Cameron Wake OLB 66 14 10.9 8.5
Clay Matthews OLB 60 13.5 9.9 6
John Abraham DE 57 13 9.4 9.5

Again we see four-of-five players regress in the following season. Interestingly, Hali was second in the league and actually underperformed based on his pressure total. The only player to again overachieve in the following season was Ware, though we did see a regression from him in the 2012 season.

While this is just a snapshot look at the data over the past three seasons, one fact rings true – sacks are difficult to come by. When we evaluate and project pass rushers, it’s unreasonable to expect most of them to overachieve. Sure, there will always be players who outkick their coverage. But to expect them to maintain this level of production from one season to the next is completely unreasonable.

As adjusted sacks display, the raw sack number can mislead fantasy owners and create expectations for future production that are unlikely to be met. Like sacks, QB pressures alone don’t tell the whole story. But when used in conjunction with sacks, the sack conversion rate is a much more effective weapon in your IDP arsenal. While a pass rusher may not be precisely at the mean conversion rate, this frequency will give you a much more realistic sense of what to expect out of your defensive ends and rush outside linebackers.

 

Jeff Ratcliffe is the Assistant Managing Editor of Pro Football Focus Fantasy. Follow him on Twitter – @JeffRatcliffe

| Director of Fantasy

Jeff Ratcliffe is the Director of Fantasy at Pro Football Focus. He produces all of our projections and is one of the most accurate rankers in the fantasy industry. Jeff also is the host of our show on SiriusXM fantasy sports radio and is one of the main hosts of our Fantasy Slant podcast.

  • Jack Burton

    Could Watt’s inside pass rush possibly explain his high rate of pressure to sack ratio? I would think pressure up the middle leads to more sacks than pressure on the outside since the QB has less time to react and its a lot harder to avoid the sack.

  • Douglas Graham

    I think Watt will break the mold. He’s a beastie who makes life miserable for Texans OL in camp.
    With Cushing back they can’t double and triple him without someone else getting loose.
    He’s stronger and quicker with dedication to film and new techniques.
    Unless some ass rolls up on him like they did Cushing last year,I think he’ll have a better year,maybe not as many sacks,but contributing to overall team sacks. He plays prepared for backside cutblocks,so they’re likely to just bounce off him.

  • Jeff Ratcliffe

    @Jack Historically, interior pass rushers record less sacks.

    Watt’s season was very special, and the numbers really suggest he won’t be able to replicate this rate of conversion. That’s not to say he won’t have another strong year, but 20+ sacks is tough to expect.