American Fottball: When to Go for Two

[Felix_Nietsche]

Um, okay, that was data for one season.
*******************************************
Um, Okay......Link to support your 50% success rate assertion?

[DCWildcat]

No data, because you can't quantify many of the intervening variables. Assuming they don't exist, however, is preposterous. That article is shockingly stupid for something in the Times. Following its logic, it is never correct to go for 2 because of EV, when there are times that it clearly is (down 2 with :01 left on the clock after scoring a touchdown, for example). Using my previous example combined with the article's logic, there would never be an appropriate time to shoot a 3 in basketball. Ever.

I would argue that Dick Vermeil's call for the Chiefs to go for 2 was a good call. His offense had been running over the defense for the entire drive, it was the end of the game, the defense was emotional distraught and physically exhausted. The chiefs' running game had been unstoppable. I believe his chances of scoring that touchdown were well over 50%, and I think anyone watching the game would agree. I think he could expect to win more long term by going for it, rather than waiting for overtime. +EV. But no data. This is why coaches are coaches.

[Felix_Nietsche]

But this is still a joke. That's like saying "well, your chance of hitting a 3 pointer is 30%, your chance of hitting a 2 is 50%, so it's more +EV to shoot 2 pointers. Therefore, you should never shoot a 3 unless you're down 3 with 1 second left in the game."
************************************************** ****
In football, there are times when it is correct to go for two points. My point is unless your team has a better than 50% sucess rate it is -EV.

Your basketball analogy does not hold water because basketball is about good shot selection. A basketball player that has a good 3-point shot may have a 40% sucess rate when they are wide open and perhaps a 20% success rate when the other team has him well covered.

[mosta]

[ QUOTE ]
This is ridiculous. There are tons of variables (momentum, defensive fatigue, offensive line skill, running back power, fullback power, etc.), that it's insane to just throw a blanket 34% chance of success down here.
************************************************** ***
OK...37% (this is for an average team)
A team that is EXCEPTIONALLY skilled at 2-point conversions might have a 44% success rate. It is still -EV.

[/ QUOTE ]

what? do you think the chance of success on a 1-point play is 100%? is it over 90%?

[DCWildcat]

Also, the EV equations given are wrong. In reality, the value we're looking for is "more" than 2.

Take your first example, a good one, where we trail 17-8. We can't just use 2 and 1 as values in an EV equation because scoring 2--which only puts us down one possession--is far more valuable than scoring 1. The values of scoring the two point conversion and scoring a field goal are really worth much different values in the contexts of the score.

If this is unconvincing, consider it this way: the fallacy with only using 1 and 2 is that if your example had been 17-6, rather than 17-8, the values of kicking a field goal vs going for 2 would be worth the same, when they clearly are not. A coach would be an idiot to go for 2 in the first example, and much less of an idiot to go for 2 in the second. Saying the values are equal is saying that he's equally an idiot in both.

[DCWildcat]

[ QUOTE ]
But this is still a joke. That's like saying "well, your chance of hitting a 3 pointer is 30%, your chance of hitting a 2 is 50%, so it's more +EV to shoot 2 pointers. Therefore, you should never shoot a 3 unless you're down 3 with 1 second left in the game."
************************************************** ****
In football, there are times when it is correct to go for two points. My point is unless your team has a better than 50% sucess rate it is -EV.

Your basketball analogy does not hold water because basketball is about good shot selection. A basketball player that has a good 3-point shot may have a 40% sucess rate when they are wide open and perhaps a 20% success rate when the other team has him well covered.

[/ QUOTE ]

By illustrating the fallacy with that basketball example, you made my point for me.

[Felix_Nietsche]

what? do you think the chance of success on a 1-point play is 100%? is it over 90%?
*************************************************
It is about 95%. Depends on the kicker. A fair amount of kickers are 100% for an entire season.

[Felix_Nietsche]

But there are situations where, I would argue, it's >60%.
************************************************** ******
60% !!! Yikes.
You must be talking about a team that has five Larry Allens( in his prime) for an offensive line. And Dan Marino as the QB. [img]/images/graemlins/smile.gif[/img]

[Felix_Nietsche]

For 4th Qtr with 13 minutes left in the game, it is less of a mistake but still a mistake.
When there is about 8 minutes to go then I say go for it.

[DCWildcat]

I guess I'm going to have to concede this one. There's no way I can quantify what I'm trying to argue, so I'm going to reluctantly defer to the data that actually exists. I've spent most of my life making fun of stupid people who think they can accruately predict things that in reality they can't, I don't want to become a hypocrite [img]/images/graemlins/smile.gif[/img]