HFBoards

Go Back   HFBoards > NHL Eastern Conference > Metropolitan Division > New York Rangers
Mobile Hockey's Future Become a Sponsor Site Rules Support Forum vBookie Page 2
Notices

The Advanced Stats Thread Episode IV: A New Hope For Advanced Stats

Closed Thread
 
Thread Tools
Old
02-15-2017, 07:14 PM
  #751
SnowblindNYR
Registered User
 
SnowblindNYR's Avatar
 
Join Date: Nov 2011
Location: Philadelphia, PA
Posts: 25,247
vCash: 500
Quote:
Originally Posted by eyjee View Post
For now, it's a secret

I believe what I'm doing is similar to K nearest neighbor
Without looking at the link, I believe that's clustering. I've learned that at some point in school. I think K Means is the one most often used, but I believe these are similar. Not sure how that would be used in this particular case though.

SnowblindNYR is offline  
Old
02-15-2017, 07:17 PM
  #752
silverfish
Not a GM
 
silverfish's Avatar
 
Join Date: Jun 2008
Location: Alphabetrium
Country: United States
Posts: 29,888
vCash: 500
Quote:
Originally Posted by SnowblindNYR View Post
Without looking at the link, I believe that's clustering. I've learned that at some point in school. I think K Means is the one most often used, but I believe these are similar. Not sure how that would be used in this particular case though.
I just need to find the right parameters. Did some minor tweaking today and it is ever so slightly better. [added average distance which seems to have helped]

silverfish is online now  
Old
02-15-2017, 07:37 PM
  #753
Mac n Gs
Gorton plz
 
Mac n Gs's Avatar
 
Join Date: Jan 2014
Country: United States
Posts: 12,147
vCash: 500
K is so much better than corsi. I know it's hard to believe, but we're actually a good hockey team. Shocking, I know.

Mac n Gs is online now  
Old
02-15-2017, 08:06 PM
  #754
silverfish
Not a GM
 
silverfish's Avatar
 
Join Date: Jun 2008
Location: Alphabetrium
Country: United States
Posts: 29,888
vCash: 500
The WOWYs for our fourth liners (Pirri, Lindberg, Fast) with and without Vesey are truly fascinating.

Would recommend taking a look. I'm on the Vesey isn't good bandwagon. He is my new Miller. My muse. My reason for studying hockey and figuring out what I'm missing if I am missing something,

Prove me wrong you ****.

silverfish is online now  
Old
02-15-2017, 08:11 PM
  #755
Mac n Gs
Gorton plz
 
Mac n Gs's Avatar
 
Join Date: Jan 2014
Country: United States
Posts: 12,147
vCash: 500
Quote:
Originally Posted by eyjee View Post
The WOWYs for our fourth liners (Pirri, Lindberg, Fast) with and without Vesey are truly fascinating.

Would recommend taking a look. I'm on the Vesey isn't good bandwagon. He is my new Miller. My muse. My reason for studying hockey and figuring out what I'm missing if I am missing something,

Prove me wrong you ****.
K, you're wrong

Is anyone telling you Vesey wasn't good during that stretch? I've said before that he didn't mesh well with those linemates, mostly because he was with a cooled off Pirri and Lindberg when he wasn't getting consistent ice time. He was also at the "college wall." The player himself admitted it.

I'm just going to keep saying that him playing well with Stepan and Nash shouldn't be an indictment. The trio works very well together because their games are complementary.

Mac n Gs is online now  
Old
02-15-2017, 08:18 PM
  #756
silverfish
Not a GM
 
silverfish's Avatar
 
Join Date: Jun 2008
Location: Alphabetrium
Country: United States
Posts: 29,888
vCash: 500
Quote:
Originally Posted by Mac n Gs View Post
K, you're wrong

Is anyone telling you Vesey wasn't good during that stretch? I've said before that he didn't mesh well with those linemates, mostly because he was with a cooled off Pirri and Lindberg when he wasn't getting consistent ice time. He was also at the "college wall." The player himself admitted it.

I'm just going to keep saying that him playing well with Stepan and Nash shouldn't be an indictment. The trio works very well together because their games are complementary.
I'm pretty sure Pirri was fine until Vesey got put on his wing. Look at his WOWYs. It's insane.

Six of Pirri's last seven games before he was benched he had Vesey as a winger. When they were on the ice together during 5v5 play, 102 minutes, they were on the ice for zero goals for and nine goals against. That's absolutely preposterous. Pirri is a 57% GF% player away from Vesey. And it's not like he has epic linemates when he's away from Vesey.

53.8% with Fast
50% with Lindberg
80% with Grabner (lol)

And I'm not convinced that Vesey is helping anyone but himself when he plays with Stepan and Nash.

And why is the rookie wall an excuse when he's on the fourth line but it's not an excuse anymore? Because he looks better on the ice. Because his Qual Teammates has gone up.

As of right now, none of the data I see on Vesey has me excited about this guy being anything but a passenger on an NHL team's middle-6 at best.

Meanwhile, the mother****ing polar opposite in this story if Pavel Buchnevich. That dude DRIVES play.

silverfish is online now  
Old
02-15-2017, 08:18 PM
  #757
SnowblindNYR
Registered User
 
SnowblindNYR's Avatar
 
Join Date: Nov 2011
Location: Philadelphia, PA
Posts: 25,247
vCash: 500
I hate how I went to grad school and paid 60k a year learned a bunch of stuff that silverfish is discussing in this thread and I forgot it all and meanwhile he's using it like candy. I want my money back.

SnowblindNYR is offline  
Old
02-15-2017, 08:20 PM
  #758
silverfish
Not a GM
 
silverfish's Avatar
 
Join Date: Jun 2008
Location: Alphabetrium
Country: United States
Posts: 29,888
vCash: 500
Quote:
Originally Posted by SnowblindNYR View Post
I hate how I went to grad school and paid 60k a year learned a bunch of stuff that silverfish is discussing in this thread and I forgot it all and meanwhile he's using it like candy. I want my money back.
Dude, I'm faking it. All of it. I took one stats class my entire life. Unless you count marketing research, then it's two. Actually, I'm going to count marketing research because we did correlations and regressions and I actually went to that class

Which is unfortunate because it's holding me back from the jobs I actually want to do.

silverfish is online now  
Old
02-15-2017, 08:41 PM
  #759
Machinehead
You needa Chytil out
 
Machinehead's Avatar
 
Join Date: Jan 2011
Location: NYC
Country: United States
Posts: 66,912
vCash: 50
Quote:
Originally Posted by Mac n Gs View Post
K is so much better than corsi. I know it's hard to believe, but we're actually a good hockey team. Shocking, I know.
It just gives me sort of a "pulled out of thin air" feeling.

Like Shea Weber. He has a great K value.

He doesn't suppress shots, he doesn't suppress shot attempts, the shots against him come from some of the highest quality locations on the ice, and over large samples, he doesn't even suppress goals.

What is he doing to have a good value? Bubblegum and unicorns?

__________________

Thank you for the fish, Dan Girardi!
Machinehead is offline  
Old
02-15-2017, 08:44 PM
  #760
Irishguy42
Registered User
 
Irishguy42's Avatar
 
Join Date: Sep 2015
Location: Bayside
Country: United States
Posts: 11,207
vCash: 500
Quote:
Originally Posted by Machinehead View Post
It just gives me sort of a "pulled out of thin air" feeling.

Like Shea Weber. He has a great K value.

He doesn't suppress shots, he doesn't suppress shot attempts, the shots against him come from some of the highest quality locations on the ice, and over large samples, he doesn't even suppress goals.

What is he doing to have a good value? Bubblegum and unicorns?
Maybe ask Manny?

Irishguy42 is online now  
Old
02-15-2017, 08:48 PM
  #761
Machinehead
You needa Chytil out
 
Machinehead's Avatar
 
Join Date: Jan 2011
Location: NYC
Country: United States
Posts: 66,912
vCash: 50
Quote:
Originally Posted by Irishguy42 View Post
Maybe ask Manny?
Most of the time when you ask you get "obviously he's doing 'something' that the traditional analysis doesn't account for."

Something. I hate that word.

So, two things now:

1)Manny and other advanced stats bloggers actually have no idea what these players are doing well.

2)That's verbatim what the "watch the gamez nerd" crew has been telling us for years, so it's not exactly what I want to hear.

Machinehead is offline  
Old
02-15-2017, 09:01 PM
  #762
Mac n Gs
Gorton plz
 
Mac n Gs's Avatar
 
Join Date: Jan 2014
Country: United States
Posts: 12,147
vCash: 500
Quote:
Originally Posted by eyjee View Post
I'm pretty sure Pirri was fine until Vesey got put on his wing. Look at his WOWYs. It's insane.

Six of Pirri's last seven games before he was benched he had Vesey as a winger. When they were on the ice together during 5v5 play, 102 minutes, they were on the ice for zero goals for and nine goals against. That's absolutely preposterous. Pirri is a 57% GF% player away from Vesey. And it's not like he has epic linemates when he's away from Vesey.

53.8% with Fast
50% with Lindberg
80% with Grabner (lol)

And I'm not convinced that Vesey is helping anyone but himself when he plays with Stepan and Nash.

And why is the rookie wall an excuse when he's on the fourth line but it's not an excuse anymore? Because he looks better on the ice. Because his Qual Teammates has gone up.

As of right now, none of the data I see on Vesey has me excited about this guy being anything but a passenger on an NHL team's middle-6 at best.

Meanwhile, the mother****ing polar opposite in this story if Pavel Buchnevich. That dude DRIVES play.
I mean the player himself admitted his conditioning was an issue and worked on improving it before and through the all star break. Let's see how he does and then try quantifying it? (I'm saying that in the most genuine way possible). I'm gonna go look into their goals against and reform my opinion and try to get back by tomorrow night. Same goes for answering you machinehead.

My goddamn cell culture partner literally ****ed up half of our cell lines and then asked me how to do the most basic thing that we've been repeating nonstop for the past month. How the **** do these people still exist in grad school, especially a really ****ing good one at that?

Mac n Gs is online now  
Old
02-16-2017, 08:14 AM
  #763
silverfish
Not a GM
 
silverfish's Avatar
 
Join Date: Jun 2008
Location: Alphabetrium
Country: United States
Posts: 29,888
vCash: 500
Quote:
Originally Posted by Machinehead View Post
Most of the time when you ask you get "obviously he's doing 'something' that the traditional analysis doesn't account for."

Something. I hate that word.

So, two things now:

1)Manny and other advanced stats bloggers actually have no idea what these players are doing well.

2)That's verbatim what the "watch the gamez nerd" crew has been telling us for years, so it's not exactly what I want to hear.
There's an entire manifest that Manny wrote on how K works and how it's useful. There's the overview blog on K on Corsica, and there's this...

http://corsica.hockey/misc/K_Manuscript.pdf

Which is way above my education level, but likely worth diving into before making any grand claims about the metric.

silverfish is online now  
Old
02-16-2017, 10:26 AM
  #764
romba
Registered User
 
Join Date: Aug 2005
Location: Brooklyn
Country: United States
Posts: 2,251
vCash: 500
Send a message via AIM to romba
Quote:
Originally Posted by aufheben View Post
I don't think it's raw, this looks like it's showing how much a part of each team's success is because of what component, and vice versa.
That's what I got from that as well. Oh and also that Washington wins at hockey.

romba is offline  
Old
02-16-2017, 12:30 PM
  #765
SnowblindNYR
Registered User
 
SnowblindNYR's Avatar
 
Join Date: Nov 2011
Location: Philadelphia, PA
Posts: 25,247
vCash: 500
Quote:
Originally Posted by eyjee View Post
I'm pretty sure Pirri was fine until Vesey got put on his wing. Look at his WOWYs. It's insane.

Six of Pirri's last seven games before he was benched he had Vesey as a winger. When they were on the ice together during 5v5 play, 102 minutes, they were on the ice for zero goals for and nine goals against. That's absolutely preposterous. Pirri is a 57% GF% player away from Vesey. And it's not like he has epic linemates when he's away from Vesey.

53.8% with Fast
50% with Lindberg
80% with Grabner (lol)

And I'm not convinced that Vesey is helping anyone but himself when he plays with Stepan and Nash.

And why is the rookie wall an excuse when he's on the fourth line but it's not an excuse anymore? Because he looks better on the ice. Because his Qual Teammates has gone up.

As of right now, none of the data I see on Vesey has me excited about this guy being anything but a passenger on an NHL team's middle-6 at best.

Meanwhile, the mother****ing polar opposite in this story if Pavel Buchnevich. That dude DRIVES play.
I really hate myself for saying this but I wonder if the NA factor helps him in the eyes of the coach. Also weirdly despite Buch being our top prospect Vesey was more high profile coming in and I wonder if that plays a role.

SnowblindNYR is offline  
Old
02-16-2017, 12:34 PM
  #766
SnowblindNYR
Registered User
 
SnowblindNYR's Avatar
 
Join Date: Nov 2011
Location: Philadelphia, PA
Posts: 25,247
vCash: 500
Quote:
Originally Posted by eyjee View Post
Dude, I'm faking it. All of it. I took one stats class my entire life. Unless you count marketing research, then it's two. Actually, I'm going to count marketing research because we did correlations and regressions and I actually went to that class

Which is unfortunate because it's holding me back from the jobs I actually want to do.
Marketing Research is absolutely a stats class. I learned some stats in MR that I didn't learn in my stats classes. For example, Chi-Squared, I learned in a more advanced undergrad stats class, but not in a beginner stats class, but I learned it in Marketing Research. Do you know how many stats classes I took if you count Marketing Research and Data Mining type stuff? I lost count. I get simple stats and correlations. But anything beyond that I either never learned or more likely don't remember. You're actually working on this stuff. If I worked on it, I'm sure I'd be fine too. I'm just too lazy to do anything beyond regression. If I were an employer doing stuff like this I'd hire you 10 times out of 10 over me.

SnowblindNYR is offline  
Old
02-16-2017, 12:53 PM
  #767
silverfish
Not a GM
 
silverfish's Avatar
 
Join Date: Jun 2008
Location: Alphabetrium
Country: United States
Posts: 29,888
vCash: 500
Quote:
Originally Posted by SnowblindNYR View Post
Marketing Research is absolutely a stats class. I learned some stats in MR that I didn't learn in my stats classes. For example, Chi-Squared, I learned in a more advanced undergrad stats class, but not in a beginner stats class, but I learned it in Marketing Research. Do you know how many stats classes I took if you count Marketing Research and Data Mining type stuff? I lost count. I get simple stats and correlations. But anything beyond that I either never learned or more likely don't remember. You're actually working on this stuff. If I worked on it, I'm sure I'd be fine too. I'm just too lazy to do anything beyond regression. If I were an employer doing stuff like this I'd hire you 10 times out of 10 over me.
Then you'd end up getting us both fired

Start working on it with me!

silverfish is online now  
Old
02-16-2017, 01:54 PM
  #768
Raspewtin
Fire Clowneault
 
Raspewtin's Avatar
 
Join Date: May 2013
Country: France
Posts: 32,940
vCash: 50
Quote:
Originally Posted by SnowblindNYR View Post
I really hate myself for saying this but I wonder if the NA factor helps him in the eyes of the coach. Also weirdly despite Buch being our top prospect Vesey was more high profile coming in and I wonder if that plays a role.
It absolutely does. Without a doubt.

__________________
If you don't utilize any form of hockey statistics or willfully ignore them, please consider never making hockey judgments again.
Raspewtin is offline  
Old
02-16-2017, 02:52 PM
  #769
silverfish
Not a GM
 
silverfish's Avatar
 
Join Date: Jun 2008
Location: Alphabetrium
Country: United States
Posts: 29,888
vCash: 500
Quote:
Originally Posted by Raspewtin View Post
It absolutely does. Without a doubt.
I don't think so. AV never hesitated with Miller. Hell, THIS season () he was bumping Kreider to the fourth line for Puempel (he did that in at least two games that I can remember).

It's just AV's super ridiculous wheel of accountability.

silverfish is online now  
Old
02-16-2017, 03:14 PM
  #770
aufheben
Global Moderator
#RIPClendening
 
aufheben's Avatar
 
Join Date: Jan 2013
Location: New Jersey
Country: United States
Posts: 32,941
vCash: 69
Quote:
Originally Posted by Machinehead View Post
It just gives me sort of a "pulled out of thin air" feeling.

Like Shea Weber. He has a great K value.

He doesn't suppress shots, he doesn't suppress shot attempts, the shots against him come from some of the highest quality locations on the ice, and over large samples, he doesn't even suppress goals.

What is he doing to have a good value? Bubblegum and unicorns?
Where are K values for individual players?

__________________

aufheben is offline  
Old
02-16-2017, 11:50 PM
  #771
SnowblindNYR
Registered User
 
SnowblindNYR's Avatar
 
Join Date: Nov 2011
Location: Philadelphia, PA
Posts: 25,247
vCash: 500
Quote:
Originally Posted by eyjee View Post
Then you'd end up getting us both fired

Start working on it with me!
Haha. I actually started a "sports analytics" blog. I messaged you about a question I had regarding R Squared and regression. I have two entries for fantasy football. One's just pivot tables. The other is just R Squared and regression. I have hockey data dating back to a couple of seasons ago when I worked on it Christmas time 2015. I just have been too lazy to do anything with it and go back to writing my blog. Frankly, it's child play compared to things like R. I'm sure I could learn it and it could help me find some sort of marketing analytics role. I think my favorite part of the whole exercise is working in excel and managing data through excel. Nothing too fancy. But, meanwhile I'm impressing my friends who don't understand regression.

SnowblindNYR is offline  
Old
02-17-2017, 10:04 AM
  #772
silverfish
Not a GM
 
silverfish's Avatar
 
Join Date: Jun 2008
Location: Alphabetrium
Country: United States
Posts: 29,888
vCash: 500
Kreider-Zibanejad-Buchnevich somehow have only played 8 mins of 5v4 time together this year.

22 shot attempts
2 goals.

This leads the team by 40 shot attempts for per 60 (small sample size alert).

Quote:
Originally Posted by SnowblindNYR View Post
Haha. I actually started a "sports analytics" blog. I messaged you about a question I had regarding R Squared and regression. I have two entries for fantasy football. One's just pivot tables. The other is just R Squared and regression. I have hockey data dating back to a couple of seasons ago when I worked on it Christmas time 2015. I just have been too lazy to do anything with it and go back to writing my blog. Frankly, it's child play compared to things like R. I'm sure I could learn it and it could help me find some sort of marketing analytics role. I think my favorite part of the whole exercise is working in excel and managing data through excel. Nothing too fancy. But, meanwhile I'm impressing my friends who don't understand regression.
Excel can get you all the way there in terms of what's available in hockey stats today. Corsica makes it easy enough to download all the data as a .CSV and manipulate it as you need to. I spend my entire work day in Excel. I love it.

R comes in handy with repeatability and sharing work and better visualizations, too. And probably more powerful statistical computations. But, if I put the time in, I'm pretty certain I'd be able to reproduce my player projection tool in Excel.

silverfish is online now  
Old
02-17-2017, 10:58 AM
  #773
ManUtdTobbe
Registered User
 
Join Date: Jun 2016
Location: Sweden
Posts: 2,227
vCash: 500
Think this is the best place for this, move it if you see fit mods.

DTMAboutHeart with a Q and A on twitter, some nyr tweets here:

"What are the New York Rangers?"
ManUtdTobbe is online now  
Old
02-17-2017, 01:43 PM
  #774
Capt McDonut
That was Vesey
 
Capt McDonut's Avatar
 
Join Date: Mar 2014
Location: florida
Country: United States
Posts: 7,271
vCash: 500
Mcdonagh is better than doughty? by what measure? please inform me good sir!

Capt McDonut is offline  
Old
02-17-2017, 04:08 PM
  #775
silverfish
Not a GM
 
silverfish's Avatar
 
Join Date: Jun 2008
Location: Alphabetrium
Country: United States
Posts: 29,888
vCash: 500
On the brightside, for my model, I just quickly mashed together the R^2 14-15 to 15-16 for goals per 60 and points per 60, and they are quite a bit lower than my model

goals per 60 YoY: r^2 .1535
primary points per 60 YoY: r^2 .2412

Right now, my model is hovering at just over 0.3 for goals, and 0.34 for primary points.

Get some.

[min 41 games both seasons]

silverfish is online now  
Closed Thread

Forum Jump


Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



All times are GMT -5. The time now is 03:59 PM.

monitoring_string = "e4251c93e2ba248d29da988d93bf5144"

Powered by vBulletin®
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
HFBoards.com is a property of CraveOnline Media, LLC, an Evolve Media, LLC company. @2017 All Rights Reserved.