There must be some research done on the EV of this somewhere??? There has to be a reason that at least some NFL coaches don't do this.

I want to say that the 49ers incessant conservatism hurt them, but they won the game so I don't really have a basis for my critism other than they most likely could of won by more. Running it on 3rd and long three times in a close game has got to hurt Alex Smith's confidence. The one fumble has been his only blatant mistake, and while it was pathetic, the dude hasn't thrown a INT yet. I really miss Norv Turner sometimes.

The reason is probably that most of them are ignorant. The one's that are aware are too old school and stubborn.

I think it's to avoid explaining it to the media if it costs you the game. If you do it and win, you're a genius. If you do it and lose, you're an even bigger idiot. Media and fans don't understand variance and results-oriented-thinking, so coaches take the safe route.

I totally agree. For some reason everyone assumes the result of the 2pt play would be different? Seems to me you're better off knowing ahead of time how many possessions you need.

You can't put that on Reid.

K 1-21 K43 Andre' Woodson pass complete to Steve Johnson for 57 yards to the LOU0,
1ST DOWN UK, TOUCHDOWN, clock 00:28.
Lones Seiber kick attempt good.

==========================
LOUISVILLE 34, KENTUCKY 40
==========================

As it turns out, they called the timeout because they only had 10 guys on the field. (And to ice the kicker).

I've tried doing this with some peers at a University widely considered to be a top notch school of high education. I've written out the math for them, tried explaining it in less complicated terms, done the whole variance shpeel, but the conversation always comes down to two points. 1) Statistics don't take into account "intangibles" like home field advantage and momentum, and 2) When it comes down to it, no coach in the NFL is going to go against conventional wisdom when their job is on the line thus they'll never do this.

I've gotten some friends to agree that in a theoretical sense going for two down 8 is correct, but they still aren't sold.

However, I think it is more correct to do it in the NFL where possession in OT is determined by a coin flop whereas in college skill level does come into play much more.

Bingo, this is totally it. The first coach that does this in a big game and it fails, will be incorrectly dubbed a moron ad nauseum by every 2 bit sports radio jerk with a microphone and every incompitent analyst at ESPN.

Going for 1 to make it 7 is "safe" and low variance and doesn't rock the boat.

I think eventually, within 10 years or so, some coach is going to come along and make correct but non-conventional wisdom and really expose some of this stuff. Then he'll have to deal with complaints from the NFLPA on behalf of the punters of the world.

Colts this week. 1st and goal at the Titans 8 with 0:18 left in the half. 4 yard completion and rather than calling a TO, they take 14 seconds to get a snap off and throw an incomplete pass with 0:01 on the clock. They are very lucky the time didn’t run out, because it was close. So now they kick the FG on 3rd down with an unused TO on the board.

This is certainly not as egregious an error as some others pointed out here, but I have Peyton as my QB and lost by 1 point this week. I’d like to have that extra shot at the end zone.

we're not talking about the average sports fan though. We're talking about a coach or coaching staff that is sharp enough to 'get it'.
I agree the pressure from media and fans is high. But I'm just saying that sooner or later it's going to have to happen somewhere. They can't keep making the same mistakes forever just because it's the old-school way of doing things. Somebody is going to get it at some point. Perhaps at the high-school or small-college level or something. But it will happen and it might just grow from there.

IMO, there are some coaches out there who are open-minded and experimental enough to try this at the NCAA level.
Not a ton of course. But there are a handful who might be smart enough to 'get it' and also be willing to implement it.

One guy I would have loved to have sat down and tried to chat about this stuff with is the late Terry Hoeppner formerly of Indiana and Miami, Ohio.
I don't know if he would have quite gotten it. But I think there was a slight chance with him.

Some of these guys are always looking at different ideas and different ways of approaching the problem. Having the 2ndary cover in a different way that can confuse the other team's QB. They meet with a couple other coaches to discuss in detail the different things they are doing.
Stuff like that.

The guys not afraid to experiment who are trying to build up some losing program like an Indiana or something are the kinds of guys who just might be able to pull off something like this.

Math geeks have flooded into baseball now and I think it has changed the game a bit as a result. The innovative GM's are actually listening to these guys. Mistakes are still being made all over the place of course but the fact is that the math people are not being completely ignored.

Football seems to be changing all the time. It's pretty different than it was 15 yrs ago or 25 yrs ago, etc etc.
At some point there will be a realization that the high chance of converting on the 2-pointer after the TD needs to be looked at much more closely.

At some point in the future such a new thing could end up being not that much different to the public than some underdog basketball team that fires up a bunch of 3-pointers to stay in a game instead of going for mostly 2-pointers.

bob - I just don't see it. The minor increase in gameEV doesn't justify the decrease in jobEV.

I don't think you'll see the breakthrough begin in the pro level. The risk is too much. But some I-AA teams in college? Sure. If that gets the ball rolling eventually you'll see it make its way up to the pro level.

The problem is, (and I'll just make up some useless random numbers to illustrate) say you have conventional wisdom Plan A, which works 45% of the time, and you have theoretically optimal Plan B, which works 55% of the time. This is a pretty significant edge, over the course of a season, but this edge is nowhere NEAR as big as the confirmation bias and recall bias that the media and the average sports fan have. You'd need a spread much larger than 10%, imo. Just look at kickers, people only remember the misses. This guy is going to have to be ridiculously successful (i.e. run really hot, because while this strategy is better it isn't WAY better) or he is going to be crucified. Then take into consideration how much he has to gain versus how much he has to lose. Play it straight, give up some edge, and you can still win, and won't be fired. Play it optimally, and while you should win a little bit more often, you are going to be fired WAY more often. It just isn't even close to worth it.

You gotta remember, the incentives for coaches are not about winning games, they are about keeping their jobs or getting better jobs. This is true of players also. For the MOST part, winning games = keeping your job, but this isn't always true. These plays are -EV for coaches, even if they are +EV for teams, because of the gross imbalance of outcomes.

Yeah, this is the succinct version of my long-ass post.

When you think about it, the job-security argument is another EV aspect too.
Most of these coaches have so little job security in the first place that they can start to think in terms of "what do they have to lose? I know it's correct and the chances of me lasting here longer than 5 years aren't that high anyway. Why not?"


I'm not saying it will happen now. Within the next 10 years MAYBE by some innovative and sharp coach of a lower big-conference school like an Indiana or Vandy or something or perhaps just some small conference.

Remember, they play football at places like Harvard and Yale and Rice and Duke and Stanford too.
Some of those players are pretty sharp and can figure this stuff out. And, as we've already noted, this really isn't THAT hard. It's not like it's advanced-calculus or something. If somebody like ME can actually get it then surely there is a coach or player SOMEWHERE who can see it too.

So some Ivy League type school has a player who actually gets it and might actually have a coach who would understand it and not be all that concerned about their job-security compared with just "lets go out there and win the damn game!!" which honestly I think is the mind-set most coaches have anyway.

Some coaches are risk-takers and love to mix it up.
Onside or fake-punt or fumblerooski where the center keeps the ball himself at weird times.
In basketball there are always 2 or 3 teams changing things pretty drastically like the Princeton slow-down offense or the Loyola-Marymount frenetic-pace.
Do you think those coaches were that concerned about job security when they implemented such game-plans and strategies that went so far against the traditional methods?

There are SOME coaches who are less risk-averse than you think and there are LOTS of coaches and teams out there.



There is that asst-coach of D1-AA James Madison who is in a wheelchair who has posted on this board before.
He, or somebody like him, would be an example of someone who understand the super-basics of game theory to perhaps implement something like this.

What you need is a combination of a coach with untouchable tenure and an innovative, open-minded approach. Its obvious why this is a rare combination. For him to have an iron-clad job with no risk of being fired, he has to have had some serious success over time. This means he will be a) old and b) confident that his system works...which it does. There is very little incentive for this guy to be innovative. Its not impossible, just rare.

Plus jocks are dumb ldo.

vhawk,

Couple of points. As mentioned before, the transition doesn't have to start at the pro level, where I agree, the money is too huge to risk getting run out of town.

Point two: baby steps. Some coaches are already smartening up about going for it in situations like 4th and 3 on the opp's 45. That can grow to drastically reducing the amount of punts attempted and it eventually becomes standard practice. It need not be one guy who comes in and always goes for two, never punts on 4th and <5, and doesn't kick field goals on 4th and <2 inside the 20.

I agree that change can be effected eventually, I'm just trying to point out that there really are some pretty fundamental barriers to these types of things, other than JUST coaches all being old, crotchety morons. They are still old crotchety morons, of course, but thats not all of it.

As a matter of fact, it would be easier to effect change from the owner's box or the GM's office, just like in baseball. Its hard because the GM has a more limited role in the NFL, but the reason sabermetrics has made such great strides in baseball is because of Beane, Theo, Ricciardi and all those guys, not because of Francona or Torre or LaRussa.

Well, that's because much (all? i'm not sure) of sabermetrics is about player evaluation, and that's the GM's job. So sure, while it may be the A's philosophy to take lots of pitches and have great plate discipline, it's largely self-fulfilled, because they go out and get guys who already do that.

Seriously, I don't understand why people have trouble with this. You HAVE to convert a 2 point conversion...your odds of converting the conversion are the same whether you do it after the first TD or the second...why would anyone wait until the second when all your eggs are in that basket? It seems like such common sense, yet I here announcers saying they should kick the extra point now and go for 2 later. What's wrong with people?

Thats not really true, there are sabermetric applications to in-game strategy, stuff like bullpen management, base stealing, sacrifices, stuff like that.

Right. The in-game stuff is relevant too.
Plus, I think that saying "they go out and get guys who already do that" is really incorrect too.
A lot of it is the instruction they receive.

If you draft a bunch of guys out of high school or college ball who are really patient at the plate but want to teach them to be super-aggressive then they are going to do that too.

Players don't just stay the same and maintain the same approach all the way through their careers. It is really a constant learning process.