In general, sack data for team defenses is not super consistent from year to year. Since 1990, the correlation coefficient between sack rate (for defenses) in Year N and sack rate in Year N+1 is 0.27. The best-fit formula (using a linear regression cover the years from ’90 to ’16) to predict sack rate for next year would be to use a constant of 4.8%, and then add 26% of the defense’s sack rate from the prior season.
That’s not too surprising of a result, but I was curious whether adding each team’s concentration index would help make sacks more predictive. As it turns out, the answer is a little complicated. I ran the same regression as above, but used each defense’s concentration index as a second variable. The change didn’t improve the correlation at all, and the p-value on the concentration index variable is 0.65, making it essentially meaningless. But it may be a little more complicated than that.
The team with the biggest decline since 1990 in sack rate, year over year, is the 2008 Chiefs. In ’07, Kansas City had a sack rate of 7%, the 8th-highest in the NFL. In 2008, it dropped to just 2%, the lowest in modern NFL history. And in 2007, Kansas City had the second most concentrated pass rush in the NFL, largely based on Jared Allen and his 15.5 sacks. In ’08, Allen was in Minnesota, and the Chiefs didn’t have a single player more than three sacks. This makes perfect sense: KC’s pass rush was very good in ’07 but centered around a superstar defender; without him the next year, the pass rush fell apart.
Sounds simple, right? Except that’s just one example. In 2000, the Titans had the 2nd best pass rush but just the 28th most concentrated: six Tennessee defenders had at least four sacks, and another six had at least two sacks, while Jevon Kearse and his 11.5 sacks made up just 21% of the team’s sacks. This sounds like a diverse pass rush that should be more sustainable from year to year, but in ’01, the team’s sack rate basically fell in half.
Analyzing sack data is very complicated: you have to factor in regression to the mean, Game Scripts, and also the randomness involved with something that only happens once every 15 or so passing plays. That said, the table below shows the 50 teams with the most concentrated pass rushes since 1982. In other words, these were the teams that were built around just one or a handful of elite pass rushers: [click to continue…]