The Advanced Stats Thread Episode IX

[Kakko]

Remember that you can use the ignore function or report button instead of responding

Previous: https://hfboards.mandatory.com/threads/the-advanced-stats-thread-episode-viii.2515233/page-41

Click to read more...

 

[Irishguy42]

Yeah, but half the fun is from responding.

 

[Machinehead]

Rangers coming for the K-rating crown.

 

[Filthy Dangles]

another bad decision by management........1000 post limit kills discussion if something gets going on in the backend of a thread. oh well....

 

[Irishguy42]

 

[SnowblindNYR]

Not sure if this should be posted here since it's a different sports, but this is the "advanced stats "thread not the "advanced stats for hockey" thread and I figured some of the advanced stats people will like this:

Here's my breakdown on the math behind going for two and why it was absolutely the right call mathematically. This is from another board.

There has been a lot written about 2 point conversions being better value than extra points. So I won't go to the general theory behind that. However, I will take a look at today's specific situation and show that Shurmur was 100% right, at least mathematically. I have a friend that argues based on qualitative factors such as momentum why he was wrong. That's an argument that if it's possible to prove is beyond my pay grade. So I will just quantify the probabilities of going for 2 and not going about it the old-fashioned way.

First of all, here are my assumptions:

1) Two TDs will be scored in regulation and the Falcons will be stopped from scoring any more points. Without that, we lose regardless of what Shurmur's call was. I guess there are wacky scenarios like a TD and 3 FGs also winning, but let's not complicate things.

2) I count ties as losses as I'm sure any head coach worth his salt does. So not winning means tying or losing.

To do the math there are really only three probabilities that need to be known. Probability of a successful PAT, probability of a successful 2 pt conversion, and the probability of winning in OT.

Probability of a successful PAT and 2 pt conversion is covered in this (unfortunately slightly outdated) FiveThirtyEight article from November 15, 2016:

According to ESPN Stats & Information Group, there have been 1,045 two-point conversion attempts since 2001,1 with teams converting 501 of those tries. That’s a 47.9 percent conversion rate; given that a successful attempt yields 2 points, that means the expected value from an average 2-point try is 0.96 points.

Interestingly, that’s almost exactly what the expected value is from an extra point these days. Since the NFL moved extra-point kicks back to the 15-yard line last season, teams have a 94.4 percent success rate, which means that an extra point has an expected value of between 0.94 and 0.95 points.

Prob of PAT: 0.944
Prob of 2pt Conversion: 0.479

For OT, I went back the last two seasons since the change to 10 minute OT. A bit of a small sample size but there have been 24 OT games and 22 of them ended with a winner. So 11/24 or ~0.458 of OT games end with a win.

Here are the following scenarios given the assumptions:

Pat Shurmur decided to go the conventional route:

That's PAT, PAT, OT.

There are two possible winning scenarios here.

Event Probability
PAT (Successful) 0.944
PAT (Successful) 0.944
Overtime (Win) 0.458333333
Prob (win) 0.408437333

Prob (win) in this scenario is calculated multiplying the probability of all of the events.

Now there's another far less likely scenario for winning. What if the first PAT is no good? The Giants still get a mulligan and can go for the 2 pt conversion to tie it and then win it in OT.

Event Probability
PAT (Unsuccessful) 0.056
2 pt Con (Successful) 0.479
Overtime (Win) 0.458333333
Prob (Win) 0.012294333

The math is done the same way. So the probability of winning with this conventional strategy is 0.408437333+0.012294333, which is about 42.07%.

Now, to do the Pat Shurmur strategy:

There are three possibilities here given the assumptions.

The 2pt conversion is good, then he goes for a PAT for the win and bipasses OT.

Event Probability
2 pt Con (Successful) 0.479
PAT (Successful) 0.944
Prob (win) 0.452176

The second scenario is the least likely one is the 2 pt conversion is good, but the PAT is missed, and then the Giants win the game in OT.

Event Probability
2 pt Con (Successful) 0.479
PAT (Unsuccessful) 0.056
Overtime (Win) 0.458333333
Prob (win) 0.012294333

The third scenario is one where the 2 pt conversion is unsuccessful, but a second 2 pt conversion is successful, and the Giants win in OT.

Event Probability
2 pt Con (Unsuccessful) 0.521
2 pt Con (Successful) 0.479
Overtime (Win) 0.458333333
Prob (win) 0.114381208

The probability of winning in this situation, if you add up all of the scenarios is ~57.9%.

Finally, a scenario dedicated to my aforementioned friend that screamed that if they were to go for 2 they should have done so after their SECOND TD, not the first. His argument was about momentum, but here I'll look at it from a mathematical perspective. I call this scenario the "reverse scenario" because it's go for 1, then go for 2, rather than the hope of going for 2 then 1.

Here there are two possible scenarios:

The ideal scenario is kicking the PAT is successful and the 2 point conversion is successful.

Event Probability
PAT (Successful) 0.944
2 pt Con (Successful) 0.479
Prob (Win) 0.452176

The other possibility is one that was already looked at. the PAT is not good, the 2 point conversion is good, and the Giants win in OT.

Event Probability
PAT (Unsuccessful) 0.056
2 pt Con (Successful) 0.479
Overtime (Win) 0.458333333
Prob (Win) 0.012294333

The total probability of the reverse strategy is ~46.4%.

Looking at this more qualitatively, this doesn't make logical sense without even looking at the numbers. In Pat Shurmur's scenario he gives you two shots at a 2 point conversion if one fails, in this scenario he gives you one shot. Another way of looking at it is that if you were to flip the scenarios. In your ideal scenario you kick the PAT and then get a 2 point conversion. Reversed is your ideal scenario of Pat Shurmur's strategy. However, if your idea scenario doesn't work and you make a PAT but miss the 2 pt conversion, you lose. On the flip side, if you miss that 2 pt conversion first, you're not going for the PAT because in that situation you KNOW it won't be enough, you'll go for a 2 pt conversion. You're basically taking away the benefit of knowing how many points will be enough to stay alive in the game go into OT.

Finally, here are the final tallies of probability of winning the game in the three strategies.

Pat Shurmur vs. Conventional Strategy +15.8%
Pat Shurmur vs. Reverse Strategy +11.4%
So mathematically, his strategy was sound.

 

[silverfish]

Can't believe the Giants are further ahead analytically than the Rangers. Loved the go for 2 call. Best thing Pat has done all season.

 

[silverfish]

Comprehensive list of players who play more 5v5 minutes per game than Pionk and also have a lower relCF% than him so far this season:

1. Libor Sulak

/end

 

[SnowblindNYR]

I posted this in the NFL thread.

Using my math the breakeven point for the probability of a successful 2 point conversion is 32.6%, in order for the call to be correct. If we look at the Giants' red zone offense as a proxy for success, as shitty as it is, it's significantly higher than that. So it every team's the last two seasons (2017 and 2018)

This is how I calculated this:

Break-even point comes when Probability of the conventional method equals the probability of Shurmur's Strategy. We're looking for a 2 point conversion percentage that makes that happen. Essentially we're solving for X.

Prob (Conventional) = Prob (Shurmur)

Prob (PAT Successful) * Prob (PAT Successful) * OT (Win) + Prob (PAT Unsuccessful) * Prob (2 pt Conversion Successful) * Prob (Overtime Win) = Prob (2 pt Conversion Successful) * Prob (PAT Successful) + Prob (2 pt Conversion Successful) * Prob (PAT Unsuccessful) * Prob (Overtime Win) + Prob (2 pt Conversion Unsuccessful) * Prob (2 pt Conversion Successful) * Prob (Overtime Win)

I plugged in the numbers we have and made the probability of getting a successful 2 pt conversion (what we're looking for) X, for easier reading.

0.4084373 + 0.025667 X = 0.944X + 0.025667X + 0.458333X * (1-X)

I won't go through every step but it gets simplified to:

0 = 0.458333X^2 - 1.402333X + 0.408437333

I have no idea how to solve that manually but I used solver in excel.

Set Objective: (Solution to the equation)

To Value of: 0

By changing variable cells: Prob (2 pt conversion)

You can also use goal seek, but it's slightly less accurate.

That spits out the answer of around. 32.6%

 

[Mac n Gs]

Can't believe the Giants are further ahead analytically than the Rangers. Loved the go for 2 call. Best thing Pat has done all season.

It's because he just copied it from the analytics stud that is Doug Pederson.

 

[silverfish]

It's because he just copied it from the analytics stud that is Doug Pederson.

That was all Namita and you know it.

 

[Fvital92]

Fast-Hayes-Chytil were injusticed out there. Better line in the game and just scored on an empty net.

 

[Blue Blooded]

Comprehensive list of players who play more 5v5 minutes per game than Pionk and also have a lower relCF% than him so far this season:

1. Libor Sulak

/end

Dafuq is a libor sulak, some sort of shoe?

 

[Blue Blooded]

The McDonagh-Strålman of 2018-19 are not the McDonagh-Strålman of 2011-14:

 

[Levitate]

Comprehensive list of players who play more 5v5 minutes per game than Pionk and also have a lower relCF% than him so far this season:

1. Libor Sulak

/end

Well here's an "interesting" thought exercise...if we're willing to say "hey this guy doesn't put up many points as a defenseman or play a traditional "defensive" style but he kills it in the advanced stats area and that shows he's an effective defensemen and should get playing time", how would we evaluate a defenseman that somehow manages to put up points (let's saying 40+ for the sake of argument) but stinks when it comes to shot metrics?

If Pionk continues to be a 5v5 black hole of shot statistics but puts up 40+ points, what do we do with that information? I guess my question is, how do we evaluate the line of where someone is more beneficial because of the points they're putting up than detrimental because of their play elsewhere.

And in more simple terms...we better hope Pionk continues to magically put up points because coaching/management sure sees him as a guy they want to ride.

e: and if this is something people have tried to quantify and I'm asking a dumb question then let me know, I'm not trying to be cute or create some kind of "gotcha" question. I have concerns about the way the team is wanting to use Pionk and his abilities out there, but 6 points in 7 games and ending the season last year on a point streak as well is...something. Putting up points because he's being played too many minutes and it's therefor not a good thing anyways? Low chance of repetition points? Not sure.

 

[DanielBrassard]

Well here's an "interesting" thought exercise...if we're willing to say "hey this guy doesn't put up many points as a defenseman or play a traditional "defensive" style but he kills it in the advanced stats area and that shows he's an effective defensemen and should get playing time", how would we evaluate a defenseman that somehow manages to put up points (let's saying 40+ for the sake of argument) but stinks when it comes to shot metrics?

If Pionk continues to be a 5v5 black hole of shot statistics but puts up 40+ points, what do we do with that information? I guess my question is, how do we evaluate the line of where someone is more beneficial because of the points they're putting up than detrimental because of their play elsewhere.

And in more simple terms...we better hope Pionk continues to magically put up points because coaching/management sure sees him as a guy they want to ride.

e: and if this is something people have tried to quantify and I'm asking a dumb question then let me know, I'm not trying to be cute or create some kind of "gotcha" question. I have concerns about the way the team is wanting to use Pionk and his abilities out there, but 6 points in 7 games and ending the season last year on a point streak as well is...something. Putting up points because he's being played too many minutes and it's therefor not a good thing anyways? Low chance of repetition points? Not sure.

Well first of all let us see if he continues to put up tons of points. But what do you do with someone who is a PP specialist who is an absolute black hole at 5v5? that's an interesting question. I guess just ask people about Shattenkirk because what he's thought of being is what Pionk has been so far.

 

[silverfish]

Well here's an "interesting" thought exercise...if we're willing to say "hey this guy doesn't put up many points as a defenseman or play a traditional "defensive" style but he kills it in the advanced stats area and that shows he's an effective defensemen and should get playing time", how would we evaluate a defenseman that somehow manages to put up points (let's saying 40+ for the sake of argument) but stinks when it comes to shot metrics?

If Pionk continues to be a 5v5 black hole of shot statistics but puts up 40+ points, what do we do with that information? I guess my question is, how do we evaluate the line of where someone is more beneficial because of the points they're putting up than detrimental because of their play elsewhere.

And in more simple terms...we better hope Pionk continues to magically put up points because coaching/management sure sees him as a guy they want to ride.

It's hard to answer this question, because no d-man is putting up 40+ points during 5v5 play, and we can't evaluate the shot metrics of the d-men who do put up 40+ points all situations because most of these guys eat PP time, they have to, to get to 40 points.

So, let's proxy. From what I remember, roughly 30 d-men a year record 40 or more points. Quick gut-check to last season shows exactly 30. The year before that, 23. I'd like to mention that it is hilarious, to me, that you think Pionk will have a 40+ point season ever in his career, but, I digress. So let's use 30 players as our proxy, and take a look at the top scoring 30 d-men last season 5v5. How many of them do you think had negative shot attempt impacts? 8. Of those 8, only 4 had an impact worse than -1. Looking at 16-17 for indexing, the same numbers are 7 and 3.

Pionk this year so far? -11.33 relCF%
Pionk last year? -6.5%

This exercise is basically to say, the guys who record a lot of points are generally good, and don't have their teams suffer the way the Rangers do, in the shot attempt game, when Pionk is on the ice.

At the end of the day, it's subjective to the user. What are you comfortable with giving up on the ice if the player is still recording points? I can only answer that for myself. When I look at it, I understand that the things Pionk are bad at are far more repeatable in the sample size we have of him than Zibanejad scoring on 2 of the 5 PP shots he's taken with Pionk on the ice so far this year.

But that's me. You can view it differently. I won't say I'm right, and I won't say you're wrong.

 

[Levitate]

But that's me. You can view it differently. I won't say I'm right, and I won't say you're wrong.

I wasn't really saying I had an opinion that was right. I guess I have a HOPE that Pionk is still young, still learning, and as the team comes together this season and future seasons with hopefully more talent and structure, that his abysmal stats even out and he can also put up respectful point totals. I hope that because from what I see this team is going to continue running him out there a lot and we have no control over his icetime so best case scenario is that he eventually starts to improve.

 

[silverfish]

I wasn't really saying I had an opinion that was right. I guess I have a HOPE that Pionk is still young, still learning, and as the team comes together this season and future seasons with hopefully more talent and structure, that his abysmal stats even out and he can also put up respectful point totals. I hope that because from what I see this team is going to continue running him out there a lot and we have no control over his icetime so best case scenario is that he eventually starts to improve.

Best case scenario is the team realizes he has a ton of shortcomings, still, isn't this all situations 1RD they like to pretend he is, and either:

1) The coaching staff fixes up what he's bad at, which is basically everything 5v5 but skating
or
2) They trade him to a team that hasn't woken up yet for a value way above his head

But, it me, silverfish, noted hater of this rebuild and a lot of the players in this organization. These posts have bias in them. I know they do.

 

[DanielBrassard]

Best case scenario is the team realizes he has a ton of shortcomings, still, isn't this all situations 1RD they like to pretend he is, and either:

1) The coaching staff fixes up what he's bad at, which is basically everything 5v5 but skating
or
2) They trade him to a team that hasn't woken up yet for a value way above his head

But, it me, silverfish, noted hater of this rebuild and a lot of the players in this organization. These posts have bias in them. I know they do.

I just think this a reasonable course of action. To expect Pionk to be a long-term fixture the way he is going right now would be a huge mistake from this FO. But they will do it.

 

[Filthy Dangles]

The McDonagh-Strålman of 2018-19 are not the McDonagh-Strålman of 2011-14:
View attachment 149283

View attachment 149285

Dan Girardi

56% CF%
4% CFRel%

Would be hillarious if this trend continued.

 

[Filthy Dangles]

Best case scenario is the team realizes he has a ton of shortcomings, still, isn't this all situations 1RD they like to pretend he is, and either:

1) The coaching staff fixes up what he's bad at, which is basically everything 5v5 but skating
or
2) They trade him to a team that hasn't woken up yet for a value way above his head

But, it me, silverfish, noted hater of this rebuild and a lot of the players in this organization. These posts have bias in them. I know they do.

I mean, what would teams realistically give up for a shot at him? A 3rd-4th round pick? I'd just rather keep him and hold onto the small hope of him being an impact player in the league.

 

[Filthy Dangles]

So......would you guys trade Pionk for a 3rd/4th? If not, do you realistically expect teams to give up something more than that for him?

 

[DanielBrassard]

I wouldn't because I think he would be valued higher than that.

 

[silverfish]