Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Teams have a less than 50% chance of making it on average

Ultimately, that's not what matters. Expected value is what matters.

Take the example from the video of a team on its own 5 yard line with 7 yards to go:

If they go for it and don't make it, they leave the other team with the ball on the 5 yard line, which they will convert into a touchdown 92% of the time, an expected value of -6.44 points. If they punt, the TD conversion goes down to 77%, an expected value of -5.39 points. Therefore, the break even is about 16% 4th down conversion rate. If a team can make it more than 1 in 6 times, they should go for it instead of punting. Since most NFL teams are in the 40-50% range, they should be going for it more often.



> Take the example from the video of a team on its own 5 yard line with 7 yards to go:

> If they go for it and don't make it, they leave the other team with the ball on the 5 yard line

Actually, if they go for it and don't make it, they either leave the other team with the ball somewhere between the goal line and just short of the 12, or they end up giving up a safety (which has both score and possession effects.)

Failing to convert on fourth down doesn't mean that you hand the ball over where you were before fourth down in the general case (only in the cases of an incomplete pass or zero-yard run, which aren't the only cases of failing to make a first down.)


Are you claiming that most NFL teams gain 7+ yards on 40-50% of plays? That can't even be close to right.


There are 18 Quarterbacks in the NFL with an average Yards per Attempt over 7 yards. So it could be in the ballpark, at least.


Average isn't relevant in this case. What you want is percent of Attempts over 7 yards (which is more akin to median).

Four zero yard plays + Four five yard plays + two 25 yard plays = 7 yards per attempt, but a 20% success rate.


Yes, but ... this discounts the practical consideration that the defense is going to change situationally, too. 4th and 2 is going to look a lot different than 4th and 17.


Yeah, I know, but I don't have those stats available. Hence why I said it's in the ballpark.


This is incorrect, because the scenario where they go for it and make it does not have a value of zero..it is still -EV


It's true, but I didn't have the data. Regardless, it decreases if they convert, which might double the break even point, but it still wouldn't be unreasonable.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: