The math behind going for it on Fourth and Short

**VeloClone** · 11-08-2012, 02:48 PM

Originally Posted by Cyclone1985

As Dial stated, it is Oregon. They are going to put up 50+ regardless.

I would like to see our 3rd down percentages VS. 4th down. Our 3rd down conversion rate is abysmal.

3rd down 39.6% (8th in league)
4th down 55.6% (7th in league)

Keep in mind almost all of ISU's 4th down attempts have been 4th and short. Not so with 3rd down attempts.

**Cyclone1985** · 11-08-2012, 02:50 PM

Originally Posted by VeloClone

3rd down 39.6
4th down 55.6

Keep in mind almost all of ISU's 4th down attempts have been 4th and short. Not so with 3rd down attempts.

No, I understand. But that is still a terrible 3rd down percentage.

**ISUME** · 11-08-2012, 03:27 PM

I think if coaches looked at yards per play more teams would go for 4th and 4 or less. Also if you know you are going to go for it on 4th down it opens up your playbook on 3rd down.

**isukendall** · 11-08-2012, 04:08 PM

I love the idea of going for it on 4th. Not that I have time to do so, but I have a strong professional background in statistical modeling. It would be very interesting to see what could be done with some stochastic modeling based on yardage per down, kick/punt return yardage, field goal %, etc. using probability distribution functions based on observed data. Would be extremely geeky, but you could objectively evaluate decisions such as going for it on 4th vs punting on a statistical basis, and see what would shake out.

Will save that for the next rainy day.....

**VeloClone** · 11-08-2012, 04:15 PM

Originally Posted by isukendall

I love the idea of going for it on 4th. Not that I have time to do so, but I have a strong professional background in statistical modeling. It would be very interesting to see what could be done with some stochastic modeling based on yardage per down, kick/punt return yardage, field goal %, etc. using probability distribution functions based on observed data. Would be extremely geeky, but you could objectively evaluate decisions such as going for it on 4th vs punting on a statistical basis, and see what would shake out.

Will save that for the next rainy day.....

Why do I have the feeling that Kirk has already done all of this?

**kberyldial** · 11-08-2012, 04:17 PM

Originally Posted by ISUME

I think if coaches looked at yards per play more teams would go for 4th and 4 or less. Also if you know you are going to go for it on 4th down it opens up your playbook on 3rd down.

The bottom line is at Oregon, they spread you out and make you account for all the skill players and speed. That forces a defense to STAY AT HOME and cover assignments. That's a big difference from an offense trying to power I a tailback up the gut for a yard on 4th down. They do what they do well. But there is no ball control factor to their game. They simply expect to outscore you. They play for **** defense. It will catch up to them against a team like Alabama - or LSU last year.

**besserheimerphat** · 11-08-2012, 04:27 PM

Originally Posted by isukendall

I love the idea of going for it on 4th. Not that I have time to do so, but I have a strong professional background in statistical modeling. It would be very interesting to see what could be done with some stochastic modeling based on yardage per down, kick/punt return yardage, field goal %, etc. using probability distribution functions based on observed data. Would be extremely geeky, but you could objectively evaluate decisions such as going for it on 4th vs punting on a statistical basis, and see what would shake out.

Will save that for the next rainy day.....

It's already been done for the NFL. In fact, at AdvancedNFLStats.com, they run a real-time Win Probability for every game, every week. You can see what plays actually made the biggest impact on Win Probability compared to what the ESPN talking heads say. Sometime's they agree, sometimes they don't. Below is the GB-JAX game last week. Mouse over the chart and it gives a description of the play. It's freakin' awesome.

Advanced NFL Stats Win Probability - Prior Weeks

**besserheimerphat** · 11-08-2012, 04:39 PM

Originally Posted by isukendall

I love the idea of going for it on 4th. Not that I have time to do so, but I have a strong professional background in statistical modeling. It would be very interesting to see what could be done with some stochastic modeling based on yardage per down, kick/punt return yardage, field goal %, etc. using probability distribution functions based on observed data. Would be extremely geeky, but you could objectively evaluate decisions such as going for it on 4th vs punting on a statistical basis, and see what would shake out.

Will save that for the next rainy day.....

It's already been done for the NFL. Seriously, check out AdvancedNFLStats.com. It's awesome. I am not affiliated with the site in any way (even if I was, it's free), but I've been reading it for a couple of years. This stuff isn't new, it's just not mainstream yet. But I'd bet that over the next decade this kind of analysis will become common-place and available down to the high school level.

AdvancedNFLStats.com has, for every game every weekend, a real-time play-by-play Win Probability calculator. They've also got a 4th down probability calculator. There's nothing available for NCAA yet, but NCAA data has only recently become readily available (check out cfbstats.com - play-by-play data for all NCAA teams from 2005 to the current week). It's only a matter of time before this kind of information is available for the NCAA. It's probably already being done by the schools themselves.