So one thing I haven't done all that clearly until now is try and quantify what exactly constitutes an elite pass rusher. I've been meaning to memorandize the way I evaluate pass rushers for some time now, so let's get to that.

Most people use sacks and quarterback hits to judge a pass rusher's productivity. And to be fair, those two stats do some heavy lifting on their own. But the raw statistics fail to capture a pass rusher's performance over the season. The raw stats don't recognize those pass rushers who scare offenses so much that opposing offenses almost always have two people blocking them in pass protection. The raw stats do not account for the defenders who create the initial pressure, thereby setting up his teammates for a hit or a sack. And the raw stats really don't count those instances when opposing quarterbacks throw a bad pass because a defender was bearing down on him. All these situations need to be factored into the equation if we are to properly differentiate the elite pass rushers from the chaff.

Professional outlets are doing a better job of trying to roll all these scenarios into the statistics, but I'll be honest - their work still leaves so much to be desired. Thankfully, the NFL is changing the game this year with NFL Pro, a publicly accessible portal showcasing some of the advanced statistics they've been promoting as part of their Next-Gen Stats mission. One incredible feature supported by NFL Pro is that it uses machine learning to track the times quarterbacks are under pressure. It notates all the plays where quarterbacks are pressured and allows users to watch those plays one after the other without any breaks. No longer do people like me have to crawl through the raw game tape to assess all the moments defenders were able to generate pressure on opposing quarterbacks - NFL Pro collates all those moments and gives me a simple interface I can use to scan through them. It's honestly beautiful. Is NFL Pro perfect in its assessment of quarterback pressures? No, of course not. But it's seriously so much better than some of these other services I see that rely on human eyes and judgement. It's made a world of difference in my research over the past month into playoff teams.

Having watched so many hours of curated tape thanks to NFL Pro, I have something resembling a formula I use to assess a pass rusher's level of work.

First, some definitions of the various types of ways quarterbacks are pressured and tailbacks can be stuffed...

• SACK - the quarterback is formally tackled before releasing the ball, thereby ending the play. Intentional grounding calls should also count toward this statistic.
• HIT - the quarterback is hit or tackled to the ground right as he releases the ball or a half-second afterwards.
• HURRY - through some action of the defense, the quarterback is forced from the safety of the pocket and/or rushed into throwing an incomplete pass because of the perception of pressure. Incomplete passes that fail in part because the quarterback cannot fully execute his throwing motion also count towards this statistic. Keep in mind that complete passes should never count toward this statistic, unless those passes fail to generate positive yardage. Scrambles that yield positive yardage do not count toward this statistic. Any time a quarterback is formally hurried, the hurry should be awarded to the nearest defender at the time.
• RUN STOP - through some action of the defense, the runner toting the ball is stopped at or behind the line of scrimmage.

Quarterback hits, hurries, sacks, and run stops can all be characterized as some form of disruption. Therefore, any defender can be mostly judged by the number of disruptions he is able to generate over time. But not all disruptions are to be treated equally. Generally put, it's all too common in the NFL to see one defender rattle the pocket and force the quarterback to retreat, only for another defender to finish the play with a hit/sack. While the end result is a single pressure, the reality is that multiple defenders were involved in the play at various points. How do we differentiate and then properly value each defender's role in pressuring the quarterback? Watching through hundreds of hours of tape, I can say one thing - the pressure from the initial defender matters so much more than any successive pressure from other defenders. The reasoning here is pretty simple - any successive pressure would never materialize without the initial pressure providing the spark.

So when it comes to pass rushing disruptions, I break them down into three categories...

• COMPLETE DISRUPTION - awarded to a defender who initiates pressure on a quarterback and then provides the final pressure that results in a sack, a hit, or a hurry.
• INITIAL DISRUPTION - awarded to a defender who initiates pressure on a quarterback, but does not provide the final pressure resulting in a sack, a hit, or a hurry. Worth more than a closing disruption.
• CLOSING DISRUPTION - awarded to a defender who does not provide the initial pressure on a quarterback, but does provide the final pressure that results in a sack, a hit, or a hurry.
• COMPLETE RUN DISRUPTION - awarded to a defender who generates a run stop all by himself without assistance from other defenders.
• ASSISTED RUN DISRUPTION - awarded to a defender who played a vital role in a run stop, but was not entirely responsible in generating that run stop.

So there we have it. We have a series of defined statistics that when properly counted can provide a very accurate picture of a pass rusher's potential. Notice that even though I emphasize the important of rushing the passer, I also incorporate a defender's ability to stop the run. Great pass rushers should also be able to limit the run to some extent.

Someday in the future, I might need to whip up some sort of AI tool that can analyze game footage and make these judgements without human assistance. I'll probably need to partner with some AI researchers to make this happen, but I know it can be done. It will probably involve spending quite a bit of time training the tool. Then again, it might be possible that the NFL will eventually release player tracking data, which means we can just write algorithms instead to assess player movements and hits. But for now, we have a way to algorithmically determine a pass rusher's worth going forward. Just how many disruptions are needed before a player is considered elite? That's a question for another time.