Games of the Week

Tonight's game notwithstanding, the game of the week was the Denver-Philadelphia game. What started to look like a blowout turned into a barnburner. Down big early, the Broncos came back to tie before surrendering a late field goal. DEN 27 PHI 30 (EI 4.9).

The runner up was the Baltimore-Pittsburgh game (4.0 EI). The Ravens were down by 10 at the half but came back to tie in the 3rd quarter. They had multiple opportunities to win the game, but imploded with penalties. BAL 20 PIT 23.

The comeback of the week was also the upset of the week. New Orleans gave up a commanding lead over Tampa Bay, and ultimately lost in overtime. TB 20 NO 17. At their low-point in the game the Buccaneers had a 3% chance of winning (CBF 33).

In last week's post at the Fifth Down I wrote about how big the Giants' 45-12 stomping of the Redskins last week ranked as the biggest blowout of the decade. There may have been bigger blowouts in terms of total points scored, but not in terms of how quickly and thoroughly one team sealed up a win. The game's EI was 0.8, which means the Redskins put up very, very little fight. Ironically, the Giants themselves were the victims of an epic blowout this week. Their 41-9 embarrassment to the Panthers only notched a 1.0 EI, ranking second in the decade to their previous game against the Redskins. It's particularly noteworthy because the Giants have been the better team than the Panthers this season and were playing at home. But even worse, it eliminated the Giants, who started 5-0, from playoff contention.

  • 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

19 Responses to “Games of the Week”

  1. Joe G says:

    Is it much effort to do a WP analysis on the Saints decision to punt the ball from their own 40 on a 4th and inches vs Tampa?

  2. Anonymous says:

    I'd love to see your analysis of Andy Reid's decision to punt the ball on 4th and 1.5 from the 50 yard line with 3 minutes left in the Eagles game. I think it was a very close call...

  3. Brian Burke says:

    Joe-You mean the one with 1:54 left in the second quarter? I show a 0.91 WP if they punt and a 0.92 WP if they go for it, a difference of only 0.01. In a vacuum, it would have technically been a good call to go for it.

    However, the Saints had such a strong upper hand at that point, there was not much room to improve their WP. The potential downside is bigger when you've already got a solid lead.

    Plus, the Saints are the stronger team, so they want low-variance outcomes. Fourth down conversion attempts are high-variance events by their nature.

  4. Brian Burke says:

    The Reid thing kills me, for 2 reasons. First, it was a repeat of the worst 4th down decision of all of 2008. Second, you don't need any fancy math or WP models to know it was not smart.

    If you have a 4th and 1 at about the 50 and you go for it, the possible outcome states are exactly equal. In other words, either I have a 1st down at the 50, or you have a 1st down at the 50. It's perfectly symmetrical.

    Plus, you know that punting will give the Broncos the advantage in the game. You don't need to know exactly how much. Just ask yourself if, with 3 minutes left in a tied game, you'd rather have the ball at your own 15 or you'd rather let them have the ball at their own 15. You'd obviously rather have the ball, (unless maybe you're the Jets).

    As long as you think your chances of getting a yard and a half is better than 50/50, it makes sense to go for it.

    As it happened (in hindsight), the punt netted 40 yards and put Denver on the 9. Very lucky for the Eagles.

    The exact numbers are huge: punting gives the Eagles an expected 0.28 WP. Going for it gives them 0.59 WP (because converting is much better than a 50/50 proposition). That's an error of 0.31. But to be honest, it's probably not quite that big because although technically it was 4th and 1, it was closer to a 4th and 2, and the conversion probability was a little lower than I gave it.

    In the end, it's possibly the worst 4th down decision of 2009.

  5. Paul says:

    "In the end, it's possible the worst 4th down decision of 2009."

    And yet, had they gone for it and missed the conversion, it would have been the Patriots all over again... But, because they went with the "safe play" no one [in the media] makes a big deal about it.

  6. Joe G says:

    Brian - I meant the 4th quarter in the TB-NO game where NO, up 17-10 had 4th and 1 (or less) on their 40 and punted to TB (which resulted in punt return for TD that tied the game).

  7. Brian Burke says:

    Joe-Ok. Gotcha. The TD return punt looks like a good decision to my system. The punt option gives NO a 0.88 WP, and the go-for-it option gives them a 0.87 WP.

    For what it's worth, the TD return was worth 0.18 WPA for TB, taking NO to a 0.70 WP. It seems like it would be a lot more catastrophic for NO, but having the ball in the final couple minutes of a tie game is worth a lot.

  8. Anonymous says:

    Back to the eagles...

    I agree with your win probability of 0.59 if they go for it (I think it should be more like 0.55, but for the most part I agree). Like you said, no fancy math needed here.

    However, I think that their win probability is much higher than 0.28 if they punt. What are your inputs for this? It doesn't make sense to me that it would be this low...

    The crowd was loud and the Broncos offense had shown an inability to move the ball all game (with the exception of points off turnovers). The Eagles have struggled in 4th and short this season. I seriously think the WP is closer to .55 than .28.

    Please show your inputs if you don't mind. I would really appreciate it and I love your columns. Keep up the great work!

  9. Dan Schlauch says:

    I think it would be worthwhile to add obvious 4th down attempts to the WP graphs. As it is now, many of the "top comebacks" are merely games in which a neccessary 4th down conversion occurred in situations where the assumed punt would have lost the game.
    It seems like you have the data to calculate 4th down conversion WPs, and it wouldn't be too hard to define a set of conditions for going for it that are Reid-proof. (for example, Losing, <2 minutes remaining)

    Also, I think there's a bug in the overtime graphs. There tend to be some pretty inexplicable spikes as one team approaches victory. Here is an obvious example:
    http://wp.advancednflstats.com/nflarchive.php?gameid=27138

  10. Brian Burke says:

    Dan-Not sure about that spike. Sometimes there is an error in the NFL play-by-play, but I'm not certain that's the case there.

    As for the obvious 4th downs, I agree. That's something I'm already working on.

  11. Brian Burke says:

    Punts from a team's own 49 typically net 33 yards, putting the Broncos on their own 18. In that situation, with 4 minutes to go, a team has about a 72% chance of winning.

    That may seem high to us, but I think this is the reason: If the Broncos can move the ball 60 yards into FG range, they can milk the clock for 3 more plays, leaving the Eagles with almost no time to respond. But if they end up punting, like they did, they still have a 0.25 WP, which equates to a 50/50 shot at stopping the Eagles and getting to OT with a 0.50 WP. And ultimately, that's what would have happened had the Broncos defense been able to make a stop on 3rd and 25.

    Yes, the Eagles aren't known for their power running game, but would that drop their true, long-term probability below 0.50? On 4th and 1 this year, they are 4 for 8, hardly enough data for a conclusion. So let's look at 3rd and 1. On 3rd and 1 they convered 16 out of 24 (67%) tries. On 3rd and 2 they converted 10 out of 15 (67%) tries. How different is 4th down than 3rd?

    Let's just be conservative and say it cost 0.20 WP instead of 0.30. Imagine being able to increase your team's chances of winning by 20 percentage points just by making a choice. It's like having Manny Mota on the bench and letting the pitcher hit.

    I may make this its own post.

  12. Marver says:

    I agree the Eagles should have certainly gone for it there. I bet the Eagles woes in short yardage situations all season had something to do with the decision.

  13. Marver says:

    Brian, before I saw those numbers I was convinced the Eagles were poor in short yardage situations. Now I think they've just been unlucky.

    The ones we remember most...
    -2nd and 1 on the New Orleans 7. McCoy losses 4 yards on a run, forcing a third down throw. Result? 97 yard INT return TD.
    -4th quarter home against Dallas: stuffed on 2nd & 1, 3rd & 1, and 4th & 1 (all runs) on the Dallas 45.
    -1st and Goal (on the 1 against San Diego) results in an incompletion and two stuffs. Reid calls for an 18 yard field goal.
    -3rd and 1 (on the 7 against San Diego). Incompletion. Akers kicks 25 yard field goal.
    -3rd and 2 (on the San Diego 25).
    McCoy losses 2 yards.

    I think when a team loses a game and the fan can go back and locate these short yardage failures, it makes us believe that the decision to go for it led to the loss. I think we should spend just as much time going back and locating short yardage conversions during wins. If we did, we wouldn't end up believing things -- "Eagles woes in short yardage situations" -- just because NFL Live talks about it every show.

  14. Anonymous says:

    how about denver punting on 4th and 2 from the 17 with like 2 min left? i have to think that was almost as bad. i mean, i'm guessing they are around 15% to win when they punt

  15. Brian Burke says:

    Yeah, it looks that way. Preliminary analysis says it was an error of about 0.10 WP, which would be the second largest of the week. I still need to firm up those numbers though. It looks like they'd have about a 0.25 WP if they punt and a 0.35 WP if they go for it.

  16. Anonymous says:

    25% seems too high to me. but yeah, i did some estimations and calculations and 15% is definitely too low. i'm guessing 21-23% is more accurate than 25%.

    i figure if the eagles can get past the 35 they are 90% to win (80% to make FG on ave). and 50% if they can't. so, to have a 75% chance to win they must get past the 35 62% of the time. i'm guessing their chances are better than that.

    keep in mind that assigning the ave net punt to denver here is generous, since punting variance favors philly. that is, a net punt of only 20 yds helps philly WAY more than a net punt of 58 yds hurts them

  17. Anonymous says:

    I thought the Broncos decision punt on 4th and 2 was simply horrendous. Punting from where they did was never going to put Eagles far outside of field goal range. In the end their D had to hold the Eagles to 7 yards of offense. Did their coach really believe they had a better chance of accomplishing that than making 4th and 2? Surely not. Just an insane decision. It was obvious they'd never get the ball back. At least if you go for it and fail the Eagles will score quickly and you get the ball back.

    No team should be punting in that situation and certainly no team with a guy who was punting as badly as the Broncos punter was.

    Just another illustration of why Belichick was absolutely right to do what he did on 4th down vs the Colts.

  18. Joe G says:

    Brian,

    Do you have any way of automatically pulling say all 4th down decisions for all teams for the entire season to determine which teams made the best "go for it or punt" decisions? Basically keep a running total of their WP for each decision. It would also tell you how much there is to gain or lose in a season from just 4th down decisions.

  19. Brian Burke says:

    Yes, I already do that. But I'm reluctant to publish it because although the automated WP model is a very good "rough draft," but it's not perfect. No model could be. I always like to confirm the numbers it spits out before I go and publish any firm conclusions.

    I'm about 75% done with a massive redo of the entire model. When I'm done, I might have enough confidence to publish stuff like that.

Leave a Reply

Note: Only a member of this blog may post a comment.