Each week during the season, I will be walking through the data from the previous Colts game analyzing the numbers to form a sort of “what happened” narrative as well as comparing the Colts against all other teams in the league. For a glossary of the stats listed, reference Season Stats. Thanks to Pro Football Reference, NFL.com, Football Outsiders and the nflFastR project for being awesome sources of weekly data.
Well, the Colts defense on the first 3 drives was impressive, but the rest of the game, not so much. Takeaways and sacks were great, but giving up 22 first downs on 28 series (78.6% DSR) is not going to win too many games.
0 EPA with 4th qtr rush success adjusted by win probability”>adj TSR,
Pen 1st/ Yds,
=20 yards or runs >= 10 yards. Not a great stat by itself but helpful in conjunction with everything else. “>Expl Plys,
19th place in PPD against ranks higher than the 24th place EPA per play against. That basically tells me that Houston left points on the field. I can think of 7 in particular.
On the season, the numbers still look good, but this defense has been trending down for a while. Since week 10 they have given up:
- the 7th most PPD
- the 3rd most yards
- the 7th highest DSR
- the 5th highest first down conversion rate
It is solely due to the 4th highest takeaway count in that period that the defense has not been much worse. If those takeaways dry up, the results will not be pretty.
= 20 yards”>20+ #/Yd
0.20 opponent EPA/db (22nd) is a far worse defensive performance than the last meeting against Deshaun Watson (0.06, 14th), as is passing success rate (59.2% vs 45.8%). In fact, pretty much all passing stats are bad and much worse than 2 weeks ago.
Watson had no problem completing passes (80.5% cmp, +16.8% cpoe) helping him move the chains on 36.7% of his dropbacks.
Again, the Colts defensive passing season stats look great, but the last 6 weeks have been pretty bad:
- 31st in PSR against
- 20th in 1st/db against
- 29th in NY/db
- 18th EPA/db
EPA isn’t as bad as the other stats due to the 4 interceptions in that time period, but if ball-hawking drops off, expect this passing defense to look like one of the worst in the league.
median EPA for similar situation, weighted by outcome (TD, first down, chunk yardage) and adjusted for 4th quarter game script by win probability. This is a pretty good predictive stat, especially since most rushing stats suck.”>adj RSR,
0. Helps explain adj RSR”>RSR,
= 10 yards”>10+ #/Yd,
Game script kept the Texans from running much, but when they did, they didn’t do well. 16.7% conversion rate and 67 total yards isn’t very efficient, which is why the Colts defense earn the #2 spot on adj. rush success rate.
Even on the ground, the Colts defense has dropped off. A season total 8th best efficiency against is tarnished by ranking only 22nd in the last 6 weeks.
CONCLUSION AND LOOK AHEAD
Whatever works, works. However, this defense has lived off of the extremes; it’s not solid play every play, it’s wildly spectacular moments that break their way. There is just no question that they have to play better if they expect any post-season success.
The good news is that the Pittsburgh offense is fading even faster. Prior to week 10, the Steelers offense ranked 9th in ppd. Since then, they are 25th (1.58) and in the last 3 weeks they are 31st (1.05). Football Outsiders ranks them 22nd in offensive DVOA but I think they are playing worse than that.
Since week 10, Roethlisberger is 22nd in efficiency (EPA/ply) and success rate. His yardage efficiency is worse, ranking 28th driven by poor accuracy ( -2.8 cpoe, 26th) and he has thrown about a pick a game.
Their run game is worse. They are ranked 30th in rushing DVOA, which matches their 6 week trailing adj RSR. Since week 10, they are 30th in first down conversion rate and dead last in EPA per carry (-0.28).
The Steeler offense is gassed. If ever there was statement game for the Colts D, this is it.