Whitney jumped ship too early

[pelagius]

Originally Posted by Indy Coug

Does he conclude that due to the law of large numbers that over the very long haul it is on average better to go for it on 4th down? With relatively few scoring plays per game, with games that are usually decided by 3 scoring plays or less, is there too much variability in his "4th down success rate" parameter to take solace in knowing poor 4th down attempts in a game that you lost will somehow pay dividends over a 25 to 50 year horizon?

I woudn't describe Romer's results that way. First, he doesn't unconditionally suggest that a team should always go for it on 4th down; there is an optimal range. Second, the benefits are large enough in terms of increased probability of winning the game that the strategy would significantly increase winning percentages over much shorting periods than 25 years. If Romer is correct, then an average team would definitely significantly increase their probability of making the playoffs.

Calibration

Romer's estimate: 4.7% per game increase in the probability winning. This is his conservative estimate. Going for it optimally in the first quarter alone increases the probability of winning by 2.3% (he actually only looks at first quarter decisions because later decisions are more likely to be affected by score considerations (e.g., one second left and a field goal will win the game)). Let's consider an average team (50% probability of going 8-8). Let's just say (for simplification) that they have a 50% of winning each game and if they follow Romer's decision rule then they would have 54.7% of winning each game. What is the probability that they go to the playoffs (so basically win 10 games or more) if they use or don't use Romer's rule?

P(win 10 or more|Don't use Romer's rule) = 22.8%
P(win 10 or more|Use Romer's Rule) = 35.7%

I think Romer's analysis has shortcomings, but if he is right then the benefits will show up over reasonable horizons.

[Insensitive PAP]

Whintey didn't actually quote the worst football coach in the history of man (Switzer) did he?

[Indy Coug]

Desperate times call for desperate measures.

[Parrot Head]

Originally Posted by pelagius

This is a fairly technical paper. Unless Bronco has been exposed to dynamic programming I am not sure how much he would understand of the paper.

Maybe that's why he hired a guy with a PhD to manage the O.