
In the 2nd quarter of the Super Bowl, Philadelphia QB Jalen Hurts throws the ball deep down the left sideline. In a tie game, Hurts was looking for a big gain to AJ Brown against Kansas City.
However, Brown was double covered, and Kansas City’s Juan Thornhill got his hands on the ball. In one possible world, the ball drops into the hands of Trent McDuffie, the other defender in coverage. Maybe he finds a seam down the sideline and scores a touchdown.
Instead, the ball fell harmlessly to the turf. While interceptions can have a massive impact on a football game, this play shows the random element in picks.
To put some numbers behind this, consider a quarterback’s interception rate, or interceptions divided by pass attempts. An NFL quarterback’s interception rate one season has little ability to predict interception rate the following season (R-squared value of 7%).
However, you can make better predictions for interceptions by expanding the set of events. As suggested earlier, the key idea is looking at plays in which a defender gets a hand on a pass.
The NFL play by play data tracks passes defended, or incomplete passes in which a defender gets a hand on the ball or jars it loose with a hit. To predict interceptions, consider bad balls, or the sum of interceptions and passes defended. These are the plays in which the QB put the ball in a dangerous position.
Bad ball rate is bad balls divided by pass attempts. From one season to the next, bad ball rate has an R-squared of 27%, about as predictive as QB metrics get.
In addition, the fraction of bad balls that turn into interceptions is random from one season to the next. There is strong regression to the NFL average of 19.7%.
If you are going to remember one thing, it should be this: quarterbacks have control over putting the football into dangerous situations. However, once they put the ball in danger, they have no control over whether the play ends as an interception.
To see the implications for the 2023 season, let’s look at New York Giants QB Daniel Jones and compare his rates with NFL averages (interception rate of 2.3%, bad ball rate of 11.6% over the past three seasons).
- 2020: interception 2.2%, bad ball 13.4%.
- 2021: interception 1.9%, bad ball 10.5%.
- 2022: interception 1.1%, bad ball 10.7%.
First, note that Jones did not excel at interception prevention because new head coach Brian Daboll showed up in 2022. Despite an injury riddled season in 2021, Jones had a better than NFL average bad ball rate.
Second, Jones will not be able to sustain a remarkable 2022 in which he only threw six interceptions. I’ve found that a three year average of bad ball rate is optimal for predicting the next season.
Over the past three seasons, Jones has had a 11.6% bad ball rate, right at NFL average. He will not be able to sustain his excellent turnover prevention in 2023.
To put this in a different way, Jones had 11.7% of his bad balls in 2022 end up as interceptions. This rate will regress to the NFL average of 19.7% in 2023.
My calculations based on bad ball rate give 10.8 interceptions for Daniel Jones this season. Although an injury can lower this estimate, it suggests value in over 8.5 interceptions (-110 available at DraftKings).
Smarter sports betting in less than 5 minutes
This article was sent to The Power Rank’s email newsletter at 5pm Eastern on Wednesday, August 23. This newsletter is a free service that strives to be:
- Valuable
- Concise
- Entertaining
Every week, I send out 5-Nugget Saturday, my curated list of sports betting tips, analytics and humor. This is the newsletter for those looking for action on any given weekend.
To get this free service delivered straight into your inbox, enter your best email address and click on “Sign up now!”
Leave a Reply