Found in the Sports Info Solutions Datahub, FPS has leveraged QB success rates relative to defensive coverage throughout the 2020 season to better predict expected QB performance relative to the market. As a matter of fact, the QB vs Coverage series was one of our best performing models of the season. The model was not as predictive in terms of team performance (wins vs. losses), but made us a lot of money on QB props, especially in the playoffs. During the season this model focused on that week’s particular matchup, QB vs. their opponent’s defense. This piece will instead look at the QB vs the aggregate, actual coverage he faced.
Here is how each QB performed by coverage, relative to their own norm (i.e. the +/- number you see along with a coverage/QB is how much better/worse that QB is in that particular set relative to HIS OWN average performance).
QB Net Success Rate by Defensive Coverage (relative to QB’s mean Success Rate). In other words, this is how much better or worse a QB ACTUALLY performs vs. certain coverages.
On a league level QB Success Rate by Coverage breaks down like this:
When looking at the NFL-wide level its quite clear that QBs, at least in 2020, preferred zone to man. Combining all zone coverages the league wide success rate is 53%, whereas league-wide man rate is 48%. On first glance, I’d suspect this is likely due to blitzing and therefore pressure being more likely in man packages. However, even when you hold for pressure rate, the zone vs man split remains the same (62% success vs zone and 57% vs man with 0 pressure).
How well a QB performs vs. a particular coverage is a fantastic tool in predicting outcomes, particularly on the micro-level. We plan to continue to find different ways to investigate the topic and leverage the data to profit on our positions.