hgfalling: hail to the juice, baby.

Jerrod Ankenman (

hgfalling) wrote,
@ 2007-11-10 18:13:00

hail to the juice, baby.
Unranked Illinois leads #1 Ohio State 28-21 with 7:30 left. They have been running the ball effectively all day. They have 4th and inches at their own thirty.

Ron Zook (the Illinois coach) sends out the punt team. As they are getting ready for the punt, Ohio State decides they have the wrong personnel out there and calls a timeout. During the timeout, Juice Williams, the Illinois quarterback, convinces Zook that they should go for it. He keeps the ball on a sneak, they make the first down, and Illinois holds the ball for the rest of the game (including three 3rd down runs for first downs by Williams) to upset OSU.

Sending out the punt team here is so wrong. When they did it I said to Michelle, "if they punt I'm going to root for Ohio State the rest of the way because any team that would punt here doesn't DESERVE to win!" Luckily, Juice fixed it.

(Post a new comment)

	terrencechan 2007-11-11 05:03 am UTC (link)
	At what distance to first down would you be indifferent to punting vs. going for it? (Reply to this)(Thread)

	hgfalling 2007-11-11 06:13 am UTC (link)
	I dunno, probably around 1.75 yards or so. (Reply to this)(Parent)

	jpmassar 2007-11-11 06:12 am UTC (link)
	Knowing virtually nothing about footballs probabilities, I am still somewhat surprised that it could be a huge mistake to punt here. What is the probability that they convert their 4th down attempt? Is the decision to go for it significantly influenced by the probability of a blocked punt? (Reply to this)(Thread)

hgfalling
2007-11-11 07:12 am UTC (link)

So I wandered around looking for 4th-and-1 conversion rates for college football, or some data to anchor to, and basically couldn't find anything, sigh. However, here's something: in the NFL in some year that we used to study it, NFL teams converted 4th and 1 68% of the time. This, of course, includes 4th and inches as well as 4th and like a whole yard, so it seems not a stretch to say that their chance of converting might be as high as 71-72% a priori, without considering that their line had been able to consistently push the OSU front back and also that their quarterback is an option quarterback, a big athlete/running back type. I'd guess their real chance of conversion was around 75%.

If you punt, as you say, the punt can get blocked, they might get a good runback, and even if they don't, you only gain some number of yards, maybe 35-40 in a good case. It's a little complicated to figure out what exactly happens if you make it or you fail, but I'll just WAG using our old ingame model:

The spread on this game was OSU -35, but given the action on the field, I think adjusting it down for model purposes to -17 seems all right.

If you go for it and fail: 67%
If you go for it and make it: 87%
Go for it total: 82%
If you punt, assume you get 35 yards net: 80%

Now our model will actually underestimate the chances that Illinois will win if they keep the ball because it doesn't mimic their clock usage until 4 minutes left or something. But it seems based on the above that it's a 2% mistake to punt.

(Reply to this)(Parent)(Thread)

tiltcity
2007-11-11 12:14 pm UTC (link)

1. If you go for it and make it, how much does it improve your chances of winning?

2. So there's some percent chance of success where the coach should be indifferent between punting and going for it; if the coach estimates that the chances of a successful 4th down conversion are less than that, he should punt.

It doesn't seem unreasonable for the coach to judge (for various reasons) that number to be smaller than the NFL percentage, even if it's not unreasonable for someone else to think that number to be higher than 71%.

(Reply to this)(Parent)

	schmengie 2007-11-11 03:42 pm UTC (link)
	http://www.econ.berkeley.edu/users/dromer/papers/PAPER_NFL_JULY05_FORWEB_CORRECTED.pdf (Reply to this)(Parent)

jpmassar
2007-11-11 05:49 pm UTC (link)

So 2% doesn't seem like a big mistake to me, and your model might not be
incorporating things that could swing it the other way such as
the wind direction (for the punt and subsequent long passes); whether the punting team has
ever had a punt blocked this season; data on the punt return stats
of the other team. Stuff a coach might conceivably be aware of
and heuristically take into account.

(Reply to this)(Parent)

	ftoc 2007-11-11 06:56 pm UTC (link)
	i like what zook said to williams: if you don't get it, i will hurt you. (Reply to this)

	barts185 2007-11-14 05:13 am UTC (link)
	And what do you make the chances of the coach getting a fair deal in the media if he goes for it, doesn't make it, and then loses the game? (Reply to this)

Username:		Create an Account Forgot your login? Login w/ OpenID
Password:
	Remember Me
	English • Español • Deutsch • Русский…