Ahem.

Freddie

Freddie deBoer used to blog at lhote.blogspot.com, and may again someday. Now he blogs here.

Related Post Roulette

7 Responses

  1. Will says:

    a) Brady’s still a dreamboat.

    b) I actually found this kind of persuasive:

    http://www.advancednflstats.com/2009/11/belichicks-4th-down-decision-vs-colts.htmlReport

    • Freddie in reply to Will says:

      But this is the problem with the metrics community entirely: they cherry pick the metric that leads to the conclusion they want. We can look at a lot of different metric that would have told the Patriots to punt. This one tells us they should have gone for it. Most advanced metrics for sports, it seems to me, are a system for finding the conclusions you want.Report

  2. sidereal says:

    We might be reading different people, but I think most of it stems from the great frustration the statistical community feels about the excessive use of the punt on 4th down. The odds say coaches punt way, way too much, and it’s rare for a coach to come along, like Belichick, who gets this and goes with the odds. So he gets a big benefit of the doubt from the sabermetrics bloggers when he does stuff like that. On the other hand, even some of the statheads think it was the wrong call.Report

  3. Michael Drew says:

    Case closed. That NYTimes Fifth Down blog post that came up at about midnight last night was particularly ridiculous.Report

  4. kth says:

    What’s odd is that this case is fairly tangential to the general case for going for it on 4th more often. Even if teams could be induced to punt less, they generally aren’t going try to convert from inside their own 30., and no one is really arguing that they should. The argument is strongest when the expected benefit from the punt is least (i.e., when there’s a good chance of a touchback).Report

  5. I think it’s a close call. The analysis that says it was clearly the statistically better decision to go for it ignores that even if Indy came back and scored, there may have been some unknown amount of time left on the clock for Tom Brady to put together a short drive to get in field goal range. As a Bills fan, I am fully aware that if Brady gets the ball back with :30 left on the clock or more, with or without timeouts, there is a 100% chance that he gets the Pats in field goal range to win the game.

    But you can’t discount the statistics, either. Even if you put them in the context of this specific game, you have to recognize that Manning still had successful touchdown drives on more than a third of his team’s possessions – and that’s without the entire field being four-down territory. Most of the statistical analyses use a figure assuming about a 40% likelihood of a touchdown after a punt based on league averages plus the Manning factor. Even if you disregard those league averages, though, and just look at the context of how this game was going (factoring in that Manning would have four downs to work with), then you still wind up with about a 40% likelihood of a touchdown after a punt.

    Keep in mind also that with the Indy offense, two minutes plus a timeout is an absolute eternity. Indy’s longest touchdown drive of the night, without Manning in the hurry-up offense, was a 90-yard drive that took up less than 3:30. His previous two touchdown drives had been a 79-yard, 2:04 drive and a 79-yard, 1:49 drive.

    It’s a coin flip of a decision. Perhaps when you have a coin flip, the best option is to take the conventional route – but that’s hardly clear.

    None of this, by the way, factors in that Tom Brady is his quarterback and probably has a higher chance of converting that fourth down than almost anyone else in the league.Report

    • Freddie in reply to Mark Thompson says:

      Actually, I think who your quarterback is on third and two should be irrelevant. If you have a power running game– that is, if you haven’t ignored your running game for at least five years in your role as GM– your QB doesn’t do anything but hand off twice and pick up the two necessary yards.Report