Will Levis was so good through the air that he broke the computers in Titans upset win over the Texans
A misunderstood stat proves just how bizarrely great Will Levis was in Week 12.
Will Levis played his best game of 2024 in Week 12 against the Houston Texans, pulling a big upset over the Tennessee Titans' division rivals.
But one metric made the rounds on social media that made it look like Levis's day wasn't just pedestrian; it made him look downright bad. The 33rd Team posted this graphic online of Week 12 QBs ranked by EPA per Dropback, and Levis is all the way down at 17th out of 25 qualifying passers.
A negative EPA per Dropback? From that game? That just doesn't sound right at all. You mean to tell me he was worse than CJ Stroud, the QB he beat and looked better than all day? Worse than Tommy DeVito? Tommy DeVito?!
Plenty of Titans fans saw this and questioned the merit of such an advanced stat when it seems so wildly misrepresentative of the game they watched with their own eyes.
And so here I am, to hopefully shed some nerdy light on why this is the way that it is. It's actually rather interesting, because what Will Levis did statistically in this game is something we almost never see:
First, a very brief baseline explanation of what the heck EPA is in the first place. EPA stands for Expected Points Added, and in the simplest possible terms, it's a way of quantifying the value of yards gained or lost in the context of points on the scoreboard.
From a handy PFF article on the topic, "Foundationally, EPA is the difference in Expected Points before and after each play. Expected points is an estimate of how many points a team will score on a drive, given the current situation (Down, Distance, Time Remaining, etc.). Intuitively, as a team gets closer to its opposing end zone, the higher the expected points of the drive."
Although it’s framed like a pure QB stat in this graphic, EPA per Dropback is an encapsulation of the efficiency of your entire offense on passing plays. So if your OL is a nightmare, or your wide receivers are dropping everything that hits them in the hands, the literal Expected Points Added (EPA!) to your scoreboard by dropping back to pass is going to be bad. That's completely regardless of how well your QB is playing.
If you’re looking for a metric that’s more agnostic to the offensive happenings around the QB, take a peek at passer rating. Against the Texans, Will Levis had a 123.3 rating, which is fantastic. That's good for 2nd best in the NFL in Week 12, which is directly at odds with his -0.09 EPA/DB (17th).
This bizarre contradiction happens whenever you're operating extremely efficiently as a passer on dropbacks when you get the pass off… but you're also being sack a trillion times.
Levis was sacked a whopping 8 times in Houston, hence the strange disconnect. Put another way, Levis was playing great football when he dropped back and didn't get sacked (rating), but he was sacked so often that the efficiency of dropbacks for the Titans on the whole was technically a slight detriment to their ability to score (EPA/DB).
This is obviously a really rare circumstance. In fact, a passer rating of 120+ and a negative EPA/DB has only happened 3 times since 2000: Jeff Garcia in 2008, Russell Wilson in 2012, and now Will Levis in 2024. And all three teams won!
So now you know. No, EPA is not a stupid, uppity advanced metric designed to tell you your eyes lied to you. It's just a tool that sometimes requires context to understand and apply!
For Titans Owner Amy Adams Strunk, the upset win over the Texans just means more in Houston
Titans and Texans ownership makes the AFC South rivalry even spicier.
