Falcons Giants Game Thread is the greatestest game thread of threads! (1 Viewer)

[Hot Sauce Guy]

No.  Odds of converting one of the two pointers is not the same as the expected value.   Extra points will still be made at .94 percent.  Two shots at .48 does not make it more likely to score two points than two extra points at .94.

Correct. 

Flip a coin 99 times. It can be heads 99x. What are the chances the coin is tails on that 100th flip? 50-50. Each 2 point attempt is a 48% chance.of success. Missing one doesn’t guarantee making the other. 

 

[joffer]

Also depends on what success rates you use.

 

[Dez]

I thought going for 2 was pretty smart there.  If you figure it is 50/50 to make it then you are giving yourself a 50% chance to win the game if you score again.  If you miss it you still have 50% chance to get the next one to tie it still.

 

[zftcg]

No.  Odds of converting one of the two pointers is not the same as the expected value.   Extra points will still be made at .94 percent.  Two shots at .48 does not make it more likely to score two points than two extra points at .94.

The EV of two two-point attempts is two points (say that five times fast). The expected value of kicking two XPs is also two.

But the more important point is that you're missing the upside of making the first one, which is what provides the 12.5% increase in W%.

 

[bb7278]

Crazy julios catch/fumble allowed him to over 100 and win countless ff matchups

 

[Ditkaless Wonders]

Correct. 

Flip a coin 99 times. It can be heads 99x. What are the chances the coin is tails on that 100th flip? 50-50. Each 2 point attempt is a 48% chance.of success. Missing one doesn’t guarantee making the other. 

Actually the smart money here would be that it is not a fair coin and to bet heads again.

 

[Hot Sauce Guy]

Actually the smart money here would be that it is not a fair coin and to bet heads again.

That’s funny you say that - it’s exactky  what smart Alec me said as a kid to my statistics teacher.

his response: No, it’s just statistically unlikely that it would come up heads 99 times in a row. That doesn’t mean it can’t happen. 

 

[zftcg]

Correct. 

Flip a coin 99 times. It can be heads 99x. What are the chances the coin is tails on that 100th flip? 50-50. Each 2 point attempt is a 48% chance.of success. Missing one doesn’t guarantee making the other. 

No one's saying it does. But if I told you I was going to flip a coin twice, and it only had to come up heads one time for you to win, you'd take the bet. Furthermore, if I told you that you would win a bonus if it came up heads the first time, with no added downside risk, you would definitely take the bet. That's this situation in a nutshell.

 

[Hot Sauce Guy]

No one's saying it does. But if I told you I was going to flip a coin twice, and it only had to come up heads one time for you to win, you'd take the bet. Furthermore, if I told you that you would win a bonus if it came up heads the first time, with no added downside risk, you would definitely take the bet. That's this situation in a nutshell.

Except you’re assuming no downside / risk. 

That’s not guaranteed. Because just like the coin could come up tails 2x, your team could miss both 2 point attempts. Easily. 

So the downside/risk is that if you miss that 1st 2 point attempt, you have a 52% chance of losing the game instead of having a 97% chance of tying it. 

 

[zftcg]

I put the Giants odds of making one of those those 2 point attempts a hell of a lot less than “the same”. 

From the Predictive Football article:

I wouldn’t argue with you that the Browns likely have a lower chance of succeeding on 2-point conversions than the average NFL team,2 which lowers their chance of winning in the go-for-it scenario. But the Browns also likely had a lower than 50% chance of winning in overtime, which has a much harsher negative effect on the win probability if you kick extra points.

Let’s estimate that the Browns only had a 35% chance of succeeding on the 2-point conversion versus the 47.5% league average. Let’s also estimate that the Browns only had a 35% chance of winning in overtime. With those estimates, going for 2 is still superior, though the benefit gained narrows to roughly 7%, or 42% versus 35%.

A good rule of thumb for this decision is that you should go for two unless you believe the gap between your probability of winning in overtime and succeeding on the 2-point conversion is more than 12%. I have trouble thinking of many cases where a team would be so much more likely to win in overtime than covert the 2. This would be limited to very poor offensive teams who still have a good chance of winning in overtime due to their strong defenses.

 

[Hot Sauce Guy]

From the Predictive Football article:

Sometimes the theory can be sound & attempting to execute it can stil be a bad idea. 

I understand the theory. I was theorize’d to death after this happened earlier this weekend. 

But that game - the way the giants were playing - the flow, & how ATL was playing defensively, it felt like the wrong call. 

Maybe the math is sound. But again - the game is played 11 on 11 between the white lines, not by a calculator. 

 

[the rover]

The EV of two two-point attempts is two points (say that five times fast). The expected value of kicking two XPs is also two.

But the more important point is that you're missing the upside of making the first one, which is what provides the 12.5% increase in W%.

Sometimes the theory can be sound & attempting to execute it can stil be a bad idea. 

I understand the theory. I was theorize’d to death after this happened earlier this weekend. 

But that game - the way the giants were playing - the flow, & how ATL was playing defensively, it felt like the wrong call. 

Maybe the math is sound. But again - the game is played 11 on 11 between the white lines, not by a calculator. 

I'll concede that the analytics call for going for it at -9 and maybe -8, depending on the time left on the clock.

The predictive football article also completely ignores ties and assumes that you're either going to win or lose in overtime, so it places a greater value on going to overtime than it should, if you value a win more than a tie or a loss.

I'm looking at a 538 article that places the increase in win probability just over 5% with 10 minutes left.   

The predictive football article also assumes that there will only be one more possession.  It doesn't take into account a result including two field goals and tying, etc.   

In any case, the Giants' coach believing he was increasing his chances of winning by 50% is nuts.

 

[zftcg]

Sometimes the theory can be sound & attempting to execute it can stil be a bad idea. 

I understand the theory. I was theorize’d to death after this happened earlier this weekend. 

But that game - the way the giants were playing - the flow, & how ATL was playing defensively, it felt like the wrong call. 

Maybe the math is sound. But again - the game is played 11 on 11 between the white lines, not by a calculator. 

See, I think that's what this comes down to. It just feels wrong, because it goes against our intuitive assumptions. So we try to come up with a bunch of post-hoc explanations for why it's wrong, or why it might be right in some circumstances, but in this case it didn't make sense (funny how people never use the in this case argument to justify a decision that goes against the conventional wisdom).

Ultimately, though, it's a truism to say that if you consistently follow the statistically recommended course of action, over the long term you will achieve the historical averages for those decisions (of course, over the long term you'll probably be fired).

Anyway, I'm probably not going to convince you tonight. But I do think the NFL will eventually have its Moneyball revolution, and teams will follow the success of Belichick and Pederson and make more analytically based (if counterintuitive) in-game decisions.

 

[zftcg]

In any case, the Giants' coach believing he was increasing his chances of winning by 50% is nuts.

I think we're currently at the stage where coaches realize they should have people crunching these numbers for them, but still don't fully understand the decisions these people are recommending. Harbaugh straight-up admitted that a few weeks ago when he said he doesn't always agree with their recommendations, but he usually follows them anyway.

My guess is that in a few years coaches will understand this stuff far more intuitively.

 

[Steeler]

It's not even that. It has nothing to do with aggression. You should always go for two in that situation. Period. It makes all the sense in the world.

I agree with you for sure - but Booger doesn't understand the math - I was saying Booger got it wrong even using his own limited context.

 

[zftcg]

Not that it matters, but I just rewatched the first 2pt attempt. It was actually a good call, they threw it to one of their two best players, and they had him open. If Eli puts it on the outside, it's a conversion and no one's arguing over Shurmur (just like no one was killing Pederson after the Eagles converted their 2pt a few weeks ago). Unfortunately, he threw it inside and the Falcons DB got a hand on it.

 

[Ditkaless Wonders]

That’s funny you say that - it’s exactky  what smart Alec me said as a kid to my statistics teacher.

his response: No, it’s just statistically unlikely that it would come up heads 99 times in a row. That doesn’t mean it can’t happen. 

While I agree that a fair coin could come up heads 99 times in a row that is not the best bet.  The best bet when a coin comes up heads 99 times in a row is that it is, in fact, not a fair coin.  Here's the kicker, if it is a fair coin you still have even odds if you go heads so it is not a wrong choice, but if as you suspect from the statistical evidence it is not a fair coin, well so much the better.

2 to the 99th power is as close to a certainty as there is for most endeavors in life.  I mean when the odds that something happens is 1 in 633825300114114700748351602688 times and it happened before you, or presuming that the game is rigged, I am going with rigged.

 

[The Frankman]

http://www.espn.com/nfl/story/_/id/25058936/pat-shurmur-new-york-giants-no-regrets-2-point-attempt-loss

The math backs up Pat Shurmur's decision to go for two down eight points with less than five minutes to go. Going for it then gives Shurmur an informational advantage. If the Giants convert, then on their next touchdown they know they only need to kick a PAT to take the lead (assuming no other scores). If they fail, which they did, they have an opportunity to go for it again to tie. The long and short of it: converting once is much more likely than failing twice.

...

The Giants ended up failing on the first two-point conversion attempt and converting the second. It proved irrelevant because the Falcons kicked a field goal in between.

At the end of the day, I still feel you guys are arguing the wrong issue. Giants ran the first two-point play, got the guy open and Eli threw a strike to OBJr. ... and he flat dropped it. I find that more troubling than anything else. Also:

Maybe the more costly late-game decisions came with less than a minute left to play. Down 23-12, they had the ball at the Atlanta 1-yard line with 45 seconds remaining. That's when Eli Manning ran a quarterback sneak and was stuffed on first down. Second down was a repeat as valuable time ticked off the clock.

The Giants were finally forced to throw on third down, when Manning connected with Odell Beckham for a touchdown with five seconds remaining. They converted the two-point conversion to make it 23-20, but without much time left.

"We got to get them in, right? We got to sneak it from the 1," Shurmur said. "I don't ... again I just saw a mush pile there, so I don't know why it didn't work. But from the 1-yard line there, we got to get it in."

Shurmur wouldn't have preferred a pass in that situation to protect from getting stuffed and the clock ticking.

"No, that's a defeatist deal," he said. "You should be able to convert on a sneak. We've all seen him do that, and for whatever reasons we didn't get it done."

After years of not being asked to do quarterback sneaks under Tom Coughlin and Ben McAdoo, that has been a staple of the Giants' offense this season. And they have been generally successful.

The weird thing is that somehow Shurmur called for Eli to sneak it... That's as odd as the playcalls in the 3rd quarter 2nd and goal.

 

[sho nuff]

http://www.espn.com/nfl/story/_/id/25058936/pat-shurmur-new-york-giants-no-regrets-2-point-attempt-loss

At the end of the day, I still feel you guys are arguing the wrong issue. Giants ran the first two-point play, got the guy open and Eli threw a strike to OBJr. ... and he flat dropped it. I find that more troubling than anything else. Also:

The weird thing is that somehow Shurmur called for Eli to sneak it... That's as odd as the playcalls in the 3rd quarter 2nd and goal.

OBJ dropped that pass no doubt...bad drop.  But I would not call that a strike.  Beckham is cutting to the outside towards the front pylon and the throw came more inside with Bekchma reaching back towards the defender.  It still got there and he needs to catch it, but it wasn't the strike or great placement IMO.

 

[Short Corner]

I'll concede that the analytics call for going for it at -9 and maybe -8, depending on the time left on the clock.

The predictive football article also completely ignores ties and assumes that you're either going to win or lose in overtime, so it places a greater value on going to overtime than it should, if you value a win more than a tie or a loss.

I'm looking at a 538 article that places the increase in win probability just over 5% with 10 minutes left.   

The predictive football article also assumes that there will only be one more possession.  It doesn't take into account a result including two field goals and tying, etc.   

In any case, the Giants' coach believing he was increasing his chances of winning by 50% is nuts.

I am guessing they consider a tie as half a win making the .5 p-value valid once the game gets to ot.

 

[Short Corner]

In any case, the Giants' coach believing he was increasing his chances of winning by 50% is nuts.

Splitting hairs here.  It was +ev and he knew it which is all that really matters.