Sunday's Division Round Analysis

Here are some quick reactions to Sunday's division round games, with at least one statistical morsel from each game.  Click the game header to see the graph and advanced box score.


The Texans outplayed the Ravens in almost every way except one--turnovers. Jacoby Jones' muffed punt was the costliest. With a 3-point lead in the 1st quarter, the HOU defense forced a 3-and-out. On the ensuing punt, Jones watched the ball bounce in front of him, but made a play for the ball to save field position. The Ravens recovered it at the HOU 6 and went on to score a TD. If you look at the WP graph, it was the signature play of the game, and HOU never fully recovered.

This game was supposed to feature two of the league's premier runners. Instead, it was all defense. Arian Foster had 132 yds on 27 carries (4.9 YPC), but netted -0.01 WPA and only 1.8 EPA. Why? Failures at critical plays. Foster was killed for a 7-yd loss by Ray Lewis on a flare pass forcing a 2nd and 17 when HOU only trailed by 4. On the next play, he was strung out of bounds for just a 1-yd gain, creating a 3rd and 16. On a 3rd and 1 at the BAL 21, Foster was stuffed for no gain. And so on.

Ray Rice did much worse. He was stuffed on several critical goal line plays. And on a 2nd and 1 to seal the win, he was stopped short of the marker. Rice finished with -0.15 WPA and -6.3 EPA. These lunatics at the Baltimore Sun who week-in and week-out clamor for more carries by Rice, who cite fallacious 'when Rice gets 25 carries the Ravens win' stats, need to get a clue. (I'm looking at you Mike Preston.) A long time ago I came to the conclusion that most sports desks at our major city papers don't have the first idea about the sport on which they claim to be so authoritative, but the Baltimore Sun is in a league of its own. Those guys seem like nice folks, but...they have a duty to educate themselves. It's their job. As it is, they're no better than the average fan with his harebrained opinions.

HOU has one of the league's best defenses, and they showed up. The BAL offensive line had an awful day. They allowed 3 sacks (plus 2 more from DBs), 3 QB Hits, 3 tackles for losses, and a bunch of critical stuffs (see above). On a 3rd and inches for the victory, they allowed a 4-yard penetration by a HOU defender, forcing a stop for no gain. Overall the BAL offensive line had a -0.20 WPA and -3.5 EPA, with just a 37.9 run SR.

With 1:59 to play in the 3rd and a 4-point lead, BAL went for it on 4th down at the goal line and failed. It was a good call. Slam dunk. They didn't get it, but the defense held, forcing a 3-and-out. The ordeal ended with excellent field position for the BAL offense and burned time off the clock along the way. The difference between a 4- and 7-point lead at that point in the game isn't very large anyway, but an 11-point lead would have been enormous, and made the 4th down attempt well worthwhile.

Your advanced stat tidbit for this game is the temperature at kickoff, which was 31 degrees. I just did some number crunching on weather effects on field goal attempts, which I'll post soon. With the numbers in hand, I had a feeling about Neil Rackers' 50-yd attempt on 4th and 6 with 5:59 to play in the 3rd quarter. As they lined up for the kick, I typed, "Too long," on the chat at the WP graph page. Sure enough, Rackers nailed it! Dead center. He got all of it. Boink. Off the crossbar. No good. 50-yd attempts are like 60-yd attempts in 30-degree weather.


Aaron Rodgers was the leading rusher, not only of his team, but of the entire game, with 66 yds. Eli Manning played masterfully, going 21/33 with 330 yds, 3 TDs, and 1 int. In more meaningful terms, Manning had 0.49 WPA and 16.6 EPA. He was on fire witha 61% SR.

The Manning to Hicks 66-yard pass was the big play of day, worth 0.16 WP. The Kuhn fumble in the 2nd quarter was devastating for GB. It was the last moment they had the upper hand in the game, never to go above 0.50 WP again.

The 23-yd run by Bradshaw just before the half looked like it set the Giants up for a long-shot FG attempt just before halftime. But the Hail Mary prayer was answered, dropping 7 points on NYG just before halftime.

Coming out of halftime, Rodgers was stripped by Umenyiora, good for 0.13 WPA. Although it feld like the Packers were never out of the game, they never really threatened again, and the Giants held on for the win.

Mr. Discount Double-Check had his worst EPA game and his 2nd-worst WPA game of the year Sunday. He totaled 0.02 WPA and -0.4 EPA, averaging just 3.9 AYPA with 264 net passing yds and an interception.

McCarthy's decision making is going to fall under scrutiny. His 4th and 5 attempt, although smart, was unsuccessful. His surprise onside attempt didn't work. Although the numbers say surprise onside attempts can be good in an expected value sense, the NYG return team appeared to be alert and ready for one.

With 4:52 to play, the Packers scored to make it a 10-point game. The Packers would need at least one onside recovery to have a prayer. McCarthy went for the onside at that point, rather than play defense for the stop, then hope to recover after a subsequent score. I thought that was smart, because you might as well find out now rather than later if the onside is going to work. You can adjust your strategy, however desperate it may be, with that knowledge in hand.

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

18 Responses to “Sunday's Division Round Analysis”

  1. Jonathan says:

    Your predictions for this Sunday's opening lines were on the spot, and closer than my own picks.

  2. Tom says:

    Rogers was robbed by his own receivers, which ruined his WPA and EPA for this one. He didn't have a bad game.

  3. Anonymous says:

    Interesting (to me anyway) thought somewhat but mostly unrelated I thought I'd share. Belichick is the one coach who seems to follow the actions your numbers bear out in every way. He trades down to the second round for value, he has a RB whos best quality is that he doesn't fumble (disregarding the importance of running and emphasising the importance of turnovers), and he doesn't get pricey free agents. He is widely acknowledged as the best head coach in the league yet people that like stats are considered "stat nerds". Hmmmmmmm.

  4. Anonymous says:

    Tom, you are correct. 8+ dropped passes.

    Brian, what do you think of assigning full blame to receivers (in term of lost WPA & EPA) when they drop passes, and to QBs when they overthrow or underthrow?


  5. TRad says:

    "Belichick is the one coach who seems to follow the actions your numbers bear out in every way."

    Brian, tell the truth: you ARE Belichick, right?

  6. Anonymous says:

    b - Not every dropped pass is 100% the receiver's fault. It could be a bad throw that the receiver barely manages to get to but can't catch - 9 times out of 10 he doesn't even touch it so it's considered a bad throw instead of a drop. But there were obviously several bad drops that were totally the receiver's fault last night. It's hard to make that judgment without looking at game film. I'm wondering if assigning about 60-70% of the blame to the receiver would work.

    - L

  7. MattieShoes says:

    I know this is nitpicking but...

    Harebrained. As in bunny rabbits.

  8. schraderfan says:

    I was actually thinking "onside kick" during Crosby's run-up before the one in the first half. The way the Giants were connecting on the deep passes I felt like if the Packers were going to stop the Giants, field position wouldn't matter that much.

    Also I have to say this somewhere... I think the fumbles made more difference than anything else in this game. Without them it would have been an even contest, much like the predictions favored the Packers by essentially home field advantage.

  9. colsbikemessenger says:

    I was stunned by the Texans uber conservative offense. Did it not seem like drive after drive they ran on first down and then often ran on second down as well?

  10. Jordan says:

    The Osi strip-sack raises an interesting issue with WPA. The WPA for a play is the difference in the Giants' win probability between the game stat prior to that play and the game state after that play. But this fails to take into account the fact that the game state changes during the course of a play. Prior to Osi's strip-sack, Jennings ran a a slant and go and Ross jumped the slant. There was no safety help, so Jennings had about 20 yards of separation and a clear path to the end zone. Maybe you can't quite count that as 7 points since Rodgers might miss the throw or Jennings might drop it, but nor can you count it as less than 5. If Osi doesn't knock the ball out of Rodgers' hand, the next game state wouldn't have been Packers' ball on the 37. It would most likely have been Packers TD.

  11. Anonymous says:

    Good point Jordan - not sure how that would be possible to take into account though

  12. Anonymous says:


    "Brian, tell the truth: you ARE Belichick, right?"

    BB and BB isn't enough of a clue? :)

  13. Brian Burke says:

    Yes, there is 'intra-play' WPA in addition to play-to-play WPA. I discussed this in a post on the biggest WPA plays in Super Bowl history a year or two ago, and specifically in context of the Eli to Tyree play. Manning was in the grasp before spinning loose. At that moment in time, the Giants' WP was much lower than it was at the beginning of the play.

  14. Anonymous says:

    @ L

    When passes are recorded as drops, it basically means it was 100%+ the receivers fault, that's why there are so few actually recorded "drops" in the stats (see If there is almost any leeway at all to blame it on the throw it is not called a drop

    That being said, I watched the game and I don't think there were 8 incompletions that would have been recorded as drops, probably more like 3-4 but that is still a huge amount. There were probably about 4 more that you would expect the GB receivers to bring in on a very regular basis (like slightly high throws where the guy starts to turn up field before catching then drops the ball). I felt like GB could have easily put up 20 points in the first half when NY was getting absolutely no pressure on Rodgers (he literally had 0 pressures in his first 20 drop backs), but the missed throw to jennings in the first and the numerous drops/pseudo drops killed them.

  15. Jonathan says:

    I think Troy Aikman and the rest of the world owe the refs an apology. The replay was the definition of "probably a fumble, but inconclusive, so the play stands."

    While everybody was staring at the Jennings's butt, his calf was clearly the first part of his body to touch the ground besides his feet. If you look at the replay closely:

    a) Jennings last has control of the football at :22. His calf is very close to the turf, but the camera's view of his calf is blocked by Webster's arm. By the time the ball is clearly out, you can just about see his calf actually making contact with the turf. Ruling on the field stands.

    There was also some controversy over the helmet-to-helmet PI. It was close, but you can see Rodgers's head kinda snap. It's hard to say for sure, but it looked like helmet-to-helmet to me.

  16. Jonathan says:

    that's the best shot I can come up with. Most people are using it do crucify Leavy, but two things are clear:

    1) While he's obviously in the midst of fumbling the ball it's not clear whether he officially lost control of the ball at that point. It looks like he's still precariously cradling it like a loaf of bread. It was probably already a fumble, but it's vaguely disputable from that angle. At best, it's the first frame in which he has lost possession.

    2) More importantly, there's no way to determine whether his lower calf is in contact with the ground. You figure it hasn't, but unless you can actually see that it's not a tackle, you can't overturn. Can't.

    I don't mind "hare-brained" opinions of establishing an identity by running 25+ times or whatever, but it truly disturbs me how impulsive people tend to be when it comes time to potentially join the angry mob.

  17. Wizard says:

    Did u reverse the plays? Didnt u say osi knocked it out after the play ? besides like u can say for zillions of situations, if the queen had ba**s she'd be king

  18. jake says:

    I agree and if it had been ruled a fumble I would have had not complaint if that had not been overturned. The video just wasn't clear either way.

Leave a Reply