Eagles / Giants Week 3

[knowyourenemy]

Can we also address how insane it is that the model recommends going for it on 4th and 2 no matter what yard line you're on with no regard for the game situation?

Very flawed model. Why is anyone citing to it?

 

[DutchBird]

For the fringe situations, the model is definitely flawed (as every statistical model is). For the not fringe situations (the center of the image - and where this 4th and 8 fell into) it has much more validity.

Also, this is a probability model - therefore it tries to weigh potential benefits and potential drawbacks based on outcome.

And - again - based on the first shots (and stills) the play to make the 1st down was actually there (Blount and maybe even Jeffery). Wentz decided to hold on for way too long.

 

[DutchBird]

Can we also address how insane it is that the model recommends going for it on 4th and 2 no matter what yard line you're on with no regard for the game situation?

Very flawed model. Why is anyone citing to it?

This model is designed to help you decide what to do in the 'dead zone' - roughly the 15-20 yards between midfield and the opponent's 35-30 yard line - where you are out of FG range yet the gain by punting it is limited.

 

[knowyourenemy]

For the fringe situations, the model is definitely flawed (as every statistical model is). For the not fringe situations (the center of the image - and where this 4th and 8 fell into) it has much more validity.

Also, this is a probability model - therefore it tries to weigh potential benefits and potential drawbacks based on outcome.

And - again - based on the first shots (and stills) the play to make the 1st down was actually there (Blount and maybe even Jeffery). Wentz decided to hold on for way too long.

1. Do you advocate going for it on 4th and 2 regardless of the spot on the field and the game situation?
2. What qualifies as a "fringe" situation? I don't see how you can just say the center of the image isn't fringe. Anywhere near a border should be considered a fringe situation. And in this case, 4th and 8 from the 43 yard line is only a few yards and/or a few yards to go from a totally different recommendation. It seems the model would have more flaws/grey area in those situations than in a situation such as 4th and 3 from your own 5 yard line (which is an obvious punting situation and any model recommending otherwise isn't worth much).
3. The model also never recommends kicking a field goal past the 38-ish yard line. I suppose we should be pissed that Pederson attempted a late game FG from well outside of that range.

Bottom line -- the model is trash. Anyone relying on it is a moron.

 

[knowyourenemy]

This model is designed to help you decide what to do in the 'dead zone' - roughly the 15-20 yards between midfield and the opponent's 35-30 yard line - where you are out of FG range yet the gain by punting it is limited.

Well, it does a pretty awful job.

 

[DutchBird]

1. Do you advocate going for it on 4th and 2 regardless of the spot on the field and the game situation?

No

2. What qualifies as a "fringe" situation? I don't see how you can just say the center of the image isn't fringe. Anywhere near a border should be considered a fringe situation. And in this case, 4th and 8 from the 43 yard line is only a few yards and/or a few yards to go from a totally different recommendation. It seems the model would have more flaws/grey area in those situations than in a situation such as 4th and 3 from your own 5 yard line (which is an obvious punting situation and any model recommending otherwise isn't worth much).

I would say with less than 2 yards to go (top of the graph), and between both 30 yard lines.

3. The model also never recommends kicking a field goal past the 38-ish yard line. I suppose we should be pissed that Pederson
attempted a late game FG from well outside of that range.

Nice deliberate misrepresentation of the situation, for this model does not even take into account the end of the game situation we were in.

 

[Iggloo]

In fairness, I suspect few statistical models say to try 61-yard field goals instead of hail mary passes. But Pederson saw something in the rookie making those long kicks in practice, and decided to go for it.

 

[knowyourenemy]

No



I would say with less than 2 yards to go (top of the graph), and between both 30 yard lines.



Nice deliberate misrepresentation of the situation, for this model does not even take into account the end of the game situation we were in.

It's not a deliberate misrepresentation of anything. The model does not factor in the game situation, which makes it more useless than it would be otherwise. I'm not sure why you're married to this chart. It's a crappy chart.

Pederson going for it on 4th down from the 43 yard line is defensible in some situations. This isn't one of them.

 

[DutchBird]

In fairness, I suspect few statistical models say to try 61-yard field goals instead of hail mary passes. But Pederson saw something in the rookie making those long kicks in practice, and decided to go for it.

Yup... because what are the odds of making it? I mean, this is the 7th longest made FG in NFL history, and that is with kickers playing in domes since forever, Mile High stadium and so on. Basically the chance of succeeding with either a Hail Mary of a FG was extremely low. Danger of the FG was a block or a return.

 

[Guy Incognito]

a NFL season doesn't hang on one game. last year i was told to "shut up" when complaining about pederson's inconsistencies after the lions loss and how it's going to cost the eagles. why was i told to shut up? because the eagles were still 3-1 and barely lost that lions game.

the eagles then went 2-8 over their next 10 games. and then it was like "well, the eagles just aren't that good." off to a 2-1 start this season and it's basically the same thing. nothing has changed.. and fans are getting excited. that is, until pederson blows calls, the eagles lose some close games and they declare we need this and that in the draft.

Since there's nothing different between this year and last year, who's facing a 10 game PED suspension?

 

[Roy Munson]

how are they going to draft Barkley if they keep winning games?

 

[Iggloo]

I was definitely sweating the possibility of a Beckham return, and so were many players, reportedly, if you look at the comments from players after the game. It's almost like they were all so fixated on that that the idea of blocking the low trajectory kick was secondary.

 

[PhillyGreen]

editing.....found conflicting links

 

[PhillyGreen]

Except that apparently the same bot says you do. Seems your shot is from three years ago.

Eagles-Giants Final Score: Observations from Philadelphia’s win over New York, 27-24

Scroll down to halfway, you see the tweet with the graph. It is not even close.

Your bot is from 2013

Found this article Should Football Teams Go For it On 4th Down? This Chart Tells You

Clicking on the link takes us to this bot. What’s the Right Call on 4th Down? Ask This Robot

The pic looks identical to the bot in the article you linked.

Scroll all the way down and you will see it is from 2013. The link I provided was for the most current bot which calls for a punt.

 

[DutchBird]

Your bot is from 2013

Found this article Should Football Teams Go For it On 4th Down? This Chart Tells You

Clicking on the link takes us to this bot. What’s the Right Call on 4th Down? Ask This Robot

The pic looks identical to the bot in the article you linked.

Scroll all the way down and you will see it is from 2013. The link I provided was for the most current bot which calls for a punt.

Might be that we both have been barking up the wrong tree... based on some comments by Doug this morning:

Why Doug Pederson went for it on 4th and 8

Suggests not going for it, win probability still over 78%. Making it, win probability jumps to 90%.

 

[PhillyGreen]

Might be that we both have been barking up the wrong tree... based on some comments by Doug this morning:

Why Doug Pederson went for it on 4th and 8

Suggests not going for it, win probability still over 78%. Making it, win probability jumps to 90%.

That has to be the dumbest reason to go for it. If he coaches using stat probabilities he is not going to last long.

 

[knowyourenemy]

So Pederson's approach is even more flawed than the methodology being touted here. I don't mind Doug going against conventional wisdom if there is support for his approach. Not sure there is any support for Doug's fuzzy math that helped him decide to go for it.

 

[mannyz]

how are they going to draft Barkley if they keep winning games?

I think that boat has sailed.

 

[knowyourenemy]

how are they going to draft Barkley if they keep winning games?

I'm still sore over the fact that he originally committed to Rutgers. Just imagine how many 6-win seasons and/or meaningless bowl appearances he could have led Rutgers to over the past couple seasons.

 

[Roy Munson]

I think that boat has sailed.