Earlier this season, I looked at interceptions by win probability. Today I want to revisit that post and update the numbers through week 7 of the 2019 season.

All data comes from Ron Yurko’s scrapr github page (note: the win probability data is definitely a little bit buggy, with the win probability reversed on pick sixes. I manually fixed those, but I didn’t otherwise mess with the data. I won’t let the perfect be the enemy of the good, though, so let’s proceed).

The graph below shows how much each offense has lost on win probability on interceptions this year on the Y-Axis, and the amount of interceptions the offense has thrown on the X-Axis. The San Francisco 49ers stand out as a pretty interesting outlier here:

The 49ers have 6 interceptions and they’ve all been meaningful… and yet, the team is still 6-0. Of course, a great defense doesn’t change the fact that Jimmy Garoppolo has thrown some bad interceptions. They have all come in close games, usually early in the game; those are two factors that tend to make interceptions pretty meaningful. Compare that to the Eagles, who have 5 interceptions, but only one has been material. Carson Went has thrown two interceptions in the 4th quarter in games where the Eagles were already trailing by more than 17 points; a third interception came on 3rd-and-8 at midfield early in the game and the interception operated like a long punt, pinning Atlanta on their own 9-yard line. A fourth interception for the Eagles came from the arm of Jake Elliott on a fake field goal; the interception was damaging, but compared to an incompletion, it didn’t have any marginal effect.

We can also look at how NFL defenses have done with interceptions. Here’s the same data as above but from the perspective of defenses. Of course, the image has to be zoomed out a bit to accommodate all 32 teams:

Here’s the same graph but without New England.

The Vikings pass defense has forced 7 interceptions through seven weeks this year, but they haven’t all been meaningful. Three of them came late in the game where Minnesota already had a 98% win probability or higher! A fourth came with 18 minutes left and Minnesota up 21-0; that bumped the Vikings win probability form 93% to 98%. The Elliott fake FG interception mentioned earlier only increased Minnesota’s win probability form 85% to 89%. In fact, all but one of the defense’s interceptions (through 7 weeks) came with the team up by double digits.