Nick Bosa is the reigning NFL sack leader and Defensive Player of the Year, but it's Myles Garrett whom we forecast to record the most sacks in 2023.
For the third year in a row we're using a statistical model to predict the number of sacks the league's best pass-rushers will record, based on past performance and their 2023 situation. We're using the same model as last year (with minor tweaks), including the following inputs:
Sack totals and rates over the past two seasons
Projected snaps, courtesy of ESPN's Mike Clay
Pass rush win rate (PRWR), which uses player tracking data from NFL Next Gen Stats
How often a player's defensive coordinator/head coach has blitzed in the past
How often opposing quarterbacks on each player's upcoming schedule get sacked
A player's team's win total, which matters because a winning team forces opponents to take greater risks and pass more
How often a player rushes from edge as opposed to the interior
To be included, players have to be defensive linemen or linebackers. They also have to have rushed the passer at least 150 times in one of the past two seasons, rushed the passer on at least 25% of their snaps last season and be projected to be play at least 250 snaps this season. Rookies were excluded. One notable exclusion as a result of the criteria is Lions pass-rusher James Houston, who recorded eight sacks on just 87 pass rushes as a rookie sixth-rounder last season.
I am pleased with the model's performance the past two seasons. While it has its misses, in a comparison to 23 season-sack props set by Caesars Sportsbook and DraftKings last July, my model recorded a mean absolute error of 4.2, versus 4.6 for the betting lines (lower is better). While it's a limited sample, it was the second year in a row the model had an advantage.
For the first time, this list is not led by T.J. Watt, who is coming off an injury-shortened 2022 campaign when he had 5.5 sacks in 10 games. Let's dive into the top 10 players in projected sacks, along with a few other notable projections. The full top 50 is below as well.