Does any coach punt from the opponent's side more than Garrett?

[FuzzyLumpkins]

You're risk averse, I get it.

Zeke ran the ball 13 times in the first half. He gained 3+ yards on 12 of those runs. That's not inconsistent.
We dominated the game but were still up only 6-0 until nearly halftime. That's the downside of the risk-averse approach: you let the other team hang around. You have to look at both sides of the coin.
Finally, sure, sometimes game and matchup conditions make the case to go the other way. But not every time, which is what you have to believe if you think coaches are not too conservative. The right answer probably isn't what the various models show. But it certainly isn't what teams are doing today. As I said, I expect we'll see something quite different 20 years from now.

I am half-assed, irrelevant risk analysis averse. I love calculated risks and think in terms of probabilities in general.

And what actually happened was what you call risk averse and while it is fun to wave your hands at your preferred outcome there is no guarantee that them going for it would result in a more favorable score. The fly in your ointment is giving them a short field should they fail and their potential to score needing only ~20 yards to get into FG range Further, the game conditions in this specific game was pretty much ideal for punting as giving them back the ball entailed severely mitigated risk.

6-0 could just as easily been 6-6 or worse. As it was we were one play away from shutting them out.

It's not like we were guaranteed a score had we converted anyway. There are risks and benefits associated with either option. All have to be considered and weighed.

I also deal in specifics. If you want to talk about whether or not in a specific game situation it makes sense to punt or go for it then great. I put zero stock in a generalized analysis when making a decision for a specific team in a specific game situation.

And what various models? I have seen instances modeled in this thread once and that is it. The data and method in that example was flawed as has been pointed out repeatedly.

 

[FuzzyLumpkins]

Yes, but we were talking about the success frequency data. 3rd down success frequency data is relevant for 4th downs.

That is only true if the probable outcomes for third down are the same as they are for fourth down. I have not seen that proven.

What it really is is a crutch to overcome the small sample sizes associated with 4th down attempts.

 

[JD_KaPow]

That is only true if the probable outcomes for third down are the same as they are for fourth down. I have not seen that proven.

What it really is is a crutch to overcome the small sample sizes associated with 4th down attempts.

There's no good reason to think they'd be significantly different except for the rarest of cases.

 

[FuzzyLumpkins]

There's no good reason to think they'd be significantly different except for the rarest of cases.

That is not how skepticism and science works. A whopper induction like that needs to be proven and there is no reason to make the assumption.

 

[JD_KaPow]

I also deal in specifics. If you want to talk about whether or not in a specific game situation it makes sense to punt or go for it then great. I put zero stock in a generalized analysis when making a decision for a specific team in a specific game situation.

We have no argument there. Obviously, for any given situation, you can and should come up with a more detailed answer based on more parameters. But the difference between the models and what coaches do is so huge that you will never escape the conclusion that teams are in general too risk averse in those situations.

And what various models? I have seen instances modeled in this thread once and that is it. The data and method in that example was flawed as has been pointed out repeatedly.

Do a Google search. Many many people have looked at this, using a variety of methodologies. The conclusion is always the same.

 

[FuzzyLumpkins]

We have no argument there. Obviously, for any given situation, you can and should come up with a more detailed answer based on more parameters. But the difference between the models and what coaches do is so huge that you will never escape the conclusion that teams are in general too risk averse in those situations.Do a Google search. Many many people have looked at this, using a variety of methodologies. The conclusion is always the same.

I did a google search and found nothing.

 

[Nightman]

Yes, but we were talking about the success frequency data. 3rd down success frequency data is relevant for 4th downs.
Look, if you simply refuse to read what I've said and look at how the calculations are done, I can't help you. This is all included in the models, as I've said and shown many many times.This is not how it works. Imagine this scenario. I offer you a bet. We flip a coin. If it comes up heads, I give you $100 (going for it and succeeding). If it comes up tails, you give me $5 (going for it and failing). You can take me up on this, or decline (punt), in which case no money changes hands. Obviously, in most cases you should take me up on my offer, EVEN THOUGH if it comes up tails, that's worse for you than declining to take the bet. It's because the reward is better than the penalty. Keeping the ball with a fresh set of downs is way more of a reward than giving up an extra 20-30 yards of field position is a penalty. So no, they don't cancel each other out. Punting is bad, going for it not making it is somewhat worse, but going for it and making it is incredibly good. It's not symmetric.Because it's completely wrong. If your probability is 50%, then half the time you're going to give the other team the ball at around their 40 instead of around their 10-20. Yeah, that's bad, but you're pretending that they'll always score form the 40 and never score from the 10, but that's not remotely close to true.And this is where I give up on this argument. You refuse to read what's been written. I have not ignored the other side of the coin, as I've said and shown over and over again, but you just keep repeating this false argument that we're not accounting for it. What happens when you go for it and fail is a fundamental part of the calculation. I'm done here.

From data I saw from the DAL/NYG game here are some expected points

NYG 1st and 10 from NYG10= -.38 epa
NYG 1st and 10 from NYG20= .28 epa
NYG 1st and 10 from NYG40= 1.6 epa

DAL 4th and 3 from NYG43= .38 epa
DAL 1st and 10 from NYG40= 2.92 epa
DAL 1st and 10 from NYG35= 3.22 epa


Getting that 1st down at the opponents 35yd is 3.22 epa .....that is more than a FG or almost guaranteed points........... a failed 4th down gives the opponent 1.6 epa ....... a risk that is worth the reward of 3.22 epa

Remember that punts can also be blocked, returned for TDs, fumbled and touch-backed at the 20yd
All these outcomes are factored in

Not quite the $100 to $5 bet but well worth the chance and it absolutely improves the chances of winning

 

[Nightman]

I did a google search and found nothing.

Search better............'Is punting worth it' brought up a dozen articles........ do your own work if you are going to yell from the peanut gallery

 

[JD_KaPow]

That is not how skepticism and science works. A whopper induction like that needs to be proven and there is no reason to make the assumption.

Now that's just silly. There are all sorts of good reasons to think that 3rd downs are good proxies for optional 4th downs. That's the null hypothesis: they're the same situation, make the yards or give up the ball.

When would 3rd downs not be good proxies? Well, 3rd and very long is often a give-up play, but that would mean that 3rd down data would underestimate the success probability, and we're not generally arguing for going for it on 4th-and-15 anyway. Occasionally, a coach will decide he's going to go for it on 4th down, which may affect his 3rd down play-calling, but that's very rare.

 

[JD_KaPow]

I did a google search and found nothing.

Let me Google that for you.

 

[FuzzyLumpkins]

Let me Google that for you.

I'm not making that argument for you and I don't allow scripts so that site doesn't even work.

 

[FuzzyLumpkins]

Search better............'Is punting worth it' brought up a dozen articles........ do your own work if you are going to yell from the peanut gallery

I was talking about 3rd down data being equivalent to 4th down data. I am familiar with the EP analysis which I am not overly fond of.

 

[JD_KaPow]

I'm not making that argument for you and I don't allow scripts so that site doesn't even work.

http://www.advancedfootballanalytics.com/index.php/home/research/game-strategy/120-4th-down-study
http://www.nber.org/papers/w9024
http://harvardsportsanalysis.org/2015/01/4th-down-strategy-punt-go-for-it-or-try-the-hard-count/
https://www.footballstudyhall.com/2013/11/15/5105958/fourth-down-pulaski-academy-kevin-kelley
And so on.

Some of these link to other things as well.

 

[FuzzyLumpkins]

From data I saw from the DAL/NYG game here are some expected points

NYG 1st and 10 from NYG10= -.38 epa
NYG 1st and 10 from NYG20= .28 epa
NYG 1st and 10 from NYG40= 1.6 epa

DAL 4th and 3 from NYG43= .38 epa
DAL 1st and 10 from NYG40= 2.92 epa
DAL 1st and 10 from NYG35= 3.22 epa


Getting that 1st down at the opponents 35yd is 3.22 epa .....that is more than a FG or almost guaranteed points........... a failed 4th down gives the opponent 1.6 epa ....... a risk that is worth the reward of 3.22 epa

Remember that punts can also be blocked, returned for TDs, fumbled and touch-backed at the 20yd
All these outcomes are factored in

Not quite the $100 to $5 bet but well worth the chance and it absolutely improves the chances of winning

What is the uncertainty? I would look for myself but I cannot for obvious reasons.

 

[FuzzyLumpkins]

http://www.advancedfootballanalytics.com/index.php/home/research/game-strategy/120-4th-down-study
http://www.nber.org/papers/w9024
http://harvardsportsanalysis.org/2015/01/4th-down-strategy-punt-go-for-it-or-try-the-hard-count/
https://www.footballstudyhall.com/2013/11/15/5105958/fourth-down-pulaski-academy-kevin-kelley
And so on.

Some of these link to other things as well.

And which one says that 4th down probable outcomes are equitable to 3rd down. The last one and the Harvard one does not even say 3rd down.

EPA is noisy and uncertain but that is a different discussion.

 

[AdamJT13]

at that position on the field yes. When you are near midfield I would punt because the way kickers are today and I do not know the NYG kicker range but lets say you go for the 4th and 2 at 43 and fail now Eli only has to drive maybe 30 yards and in field goal range. most NFL offenses can put stuff together for 30 yards

The difference between giving them the ball at the 43 and giving them the ball at the 9 (0.98 points) is less than one-third of the difference between us having the ball at their 40 and us not having the ball at all (3.13 points). In the long run, you need to convert on only 20 percent of those fourth-down attempts to come out ahead. But as has been mentioned throughout this thread, NFL teams typically ignore long-term risk-reward stats because there are only 16 games, and one failure could be construed as catastrophic.

 

[xwalker]

I am half-assed, irrelevant risk analysis averse. I love calculated risks and think in terms of probabilities in general.

And what actually happened was what you call risk averse and while it is fun to wave your hands at your preferred outcome there is no guarantee that them going for it would result in a more favorable score. The fly in your ointment is giving them a short field should they fail and their potential to score needing only ~20 yards to get into FG range Further, the game conditions in this specific game was pretty much ideal for punting as giving them back the ball entailed severely mitigated risk.

6-0 could just as easily been 6-6 or worse. As it was we were one play away from shutting them out.

It's not like we were guaranteed a score had we converted anyway. There are risks and benefits associated with either option. All have to be considered and weighed.

I also deal in specifics. If you want to talk about whether or not in a specific game situation it makes sense to punt or go for it then great. I put zero stock in a generalized analysis when making a decision for a specific team in a specific game situation.

And what various models? I have seen instances modeled in this thread once and that is it. The data and method in that example was flawed as has been pointed out repeatedly.

Fuzzy earned the triple hammer.

 

[AdamJT13]

That question has absolutely nothing to do with the conversation.

Yes it does. You claim that field position is "far more important" than scoring points. Tell that to any coach whose team lost a game but had better field position in the game. Points are obviously more important than field position. Field position is important ONLY in relation to the odds of scoring (or allowing) points from that spot -- because points are what matters, not field position.

Given the scenario that the Cowboys are outside of field goal range, but are in the oppositions territory, in most cases, the smart move for the Cowboys in particular is to punt.

...

Had the Cowboys gone for it and failed to convert, the Giants would then only have to drive 15 to 20 yards to be in field goal range.

So if we're on the Giants' 43 with one of the best kickers in NFL history, we're "outside of field goal range," but the Giants have to drive only to our 42 to be in field goal range with some guy who had never made an NFL roster before? Really? You think Aldrick Rosas would have kicked a 60-yard field goal?

The Cowboys rightfully placed more faith in their offenses ability to drive up the field than their defenses ability to prevent the Giants offense from getting 15 yards. It really is that simple.

They don't have faith in the offense's ability to gain 2 yards?

 

[AdamJT13]

... it seems you are the one who is trusting your gut....go for it, because a geek in an article said or the other article that a BOT calculated its better to risk it, but then the stats show no NFL coach does it...yeah I trust a bot and a geek, but shouldn't trust Bilicheck....

so tell me genius... is Bilicheck wrong in your opinion for not going for it on 4th and 2 inside opponents territory? are you calling Billichek not a good coach?

Um, since Belichick became the Patriots' head coach in 2000, the Patriots have the HIGHEST percentage of times going for it on fourth-and-short between midfield and the opponent's 30 -- he has gone for it 64.2 percent of the time.

 

[jday]

Yes it does. You claim that field position is "far more important" than scoring points. Tell that to any coach whose team lost a game but had better field position in the game. Points are obviously more important than field position. Field position is important ONLY in relation to the odds of scoring (or allowing) points from that spot -- because points are what matters, not field position.

You sir are grossly taking what I said out of context and as a result we are having two completely different conversations. When you have a lead, the opposing offense is inept, your offense is moving up and down the field with relative ease, your defense is not your strongest side of the ball, it is fourth and 2, just outside of field goal range and beyond the oppositions 50, it is better to punt rather than go for it. If in Field Goal range given the above circumstances, get the points.

I also said in one of my earlier post that overall circumstance dictates. If we are facing a really good offensive team and it looks like it may be a shoot out, I might be a little more aggresive in that game for what should be obvious reasons. In the case of the Cowboys this past weekend however, the Cowboys knew they had a good gameplan on defense against a sputtering Giants offense and knew there was no reason to be aggressive on offense because they had confidence in their offense's ability to get back there again. The coaches did not, however, have the same level of confidence that their defense could prevent the Giants from getting within field goal range, where they might nullify one of our scores and get some momentum going in their favor.

So if we're on the Giants' 43 with one of the best kickers in NFL history, we're "outside of field goal range," but the Giants have to drive only to our 42 to be in field goal range with some guy who had never made an NFL roster before? Really? You think Aldrick Rosas would have kicked a 60-yard field goal?

If I say somewhere beyond the 50 but outside of field goal range, that means literally anywhere between. Like, it could be 49 even, cause that yard marker happens somewhere between 50 and outside of field goal range. So, let's say the Cowboys go for it on there own 49. It's fourth and 2. The opposition offense is struggling to get out of their own way. Your offense keeps moving the ball up the field, chewing up clock and knocking down field goals like bowling pins. Why in the hell wouldn't you punt it? Not converting that fourth is like breathing life into their entire team. Or do you dismiss this notion of momentum in football games?

They don't have faith in the offense's ability to gain 2 yards?

It's not a question of trust; it's a question of whether or not you take calculated risk or shoot from the hip with millions and millions of dollars on the line. To be honest, I struggle to understand why I am even debating a point in hindsight when the former strategy yielded the Cowboys a 19 to 3 victory this past weekend.