2023-24 Roster Thread #7: A shave and a haircut, two bits

[VladDrag]

Thank you. This is great. A lot to chew on, but a good intro. Is there a good place you can recommend to see a lot of these definitions side by side? or are they all basically proprietary things that everyone does differently?

So many of the calculated numbers (GAR/xGAR, RAPM-type numbers, expected goals, venue and score adjusted numbers) are specific to each reporting site, but they all come from the same raw dataset. The Real Time Scoring System is a dataset generated by the NHL that records select on ice actions. They record the time, the players on ice, the geographical location, etc. Using this data, all xGAR, xG, RAPM-type numbers are generated. This is know as a public model.

There private firms that generate their own datasets, but those are very expensive and usually only sell their data to professional sports team.

In terms of the actual terms used (Corsi, Fenwich, on-ice, etc.) this is a great starting point: Evolving-Hockey

At the end of this they link an article written by Charlie that explains more of the analytical side of things.

 

[JojoTheWhale]

Thank you. This is great. A lot to chew on, but a good intro. Is there a good place you can recommend to see a lot of these definitions side by side? or are they all basically proprietary things that everyone does differently?

There are beginners guides that explain concepts like Corsi and xGF (Hockey Advanced Analytics: What Are They & Why Are They Important?) then there are dense mathematical explanations of individual metrics. What doesn’t exist to the best of my knowledge is the broad description of how and why you might prefer X stat for Y player. The closest you can come is going to an individual site and reading how their own metrics work — MoneyPuck.com -About and How it Works for example.

 

[Beef Invictus]

So many of the calculated numbers (GAR/xGAR, RAPM-type numbers, expected goals, venue and score adjusted numbers) are specific to each reporting site, but they all come from the same raw dataset. The Real Time Scoring System is a dataset generated by the NHL that records select on ice actions. They record the time, the players on ice, the geographical location, etc. Using this data, all xGAR, xG, RAPM-type numbers are generated. This is know as a public model.

There private firms that generate their own datasets, but those are very expensive and usually only sell their data to professional sports team.

In terms of the actual terms used (Corsi, Fenwich, on-ice, etc.) this is a great starting point: Evolving-Hockey

At the end of this they link an article written by Charlie that explains more of the analytical side of things.

Always remember that Charlie knows better and is choosing to be deceptive in defense of the team

 

[Striiker]

So many of the calculated numbers (GAR/xGAR, RAPM-type numbers, expected goals, venue and score adjusted numbers) are specific to each reporting site, but they all come from the same raw dataset. The Real Time Scoring System is a dataset generated by the NHL that records select on ice actions. They record the time, the players on ice, the geographical location, etc. Using this data, all xGAR, xG, RAPM-type numbers are generated. This is know as a public model.

There private firms that generate their own datasets, but those are very expensive and usually only sell their data to professional sports team.

In terms of the actual terms used (Corsi, Fenwich, on-ice, etc.) this is a great starting point: Evolving-Hockey

At the end of this they link an article written by Charlie that explains more of the analytical side of things.

Which was the one that said Ghost was the best defensemen ever, because that's the one I think is most accurate.

 

[chadateit]

There are beginners guides that explain concepts like Corsi and xGF (Hockey Advanced Analytics: What Are They & Why Are They Important?) then there are dense mathematical explanations of individual metrics. What doesn’t exist to the best of my knowledge is the broad description of how and why you might prefer X stat for Y player. The closest you can come is going to an individual site and reading how their own metrics work — MoneyPuck.com -About and How it Works for example.

Awesome. Thank you. I'm a math guy, so I'm not afraid of the dry math side of things, but need to understand the qualitative part of it. I've been relying on you guys for too long. About time to start learning.

So many of the calculated numbers (GAR/xGAR, RAPM-type numbers, expected goals, venue and score adjusted numbers) are specific to each reporting site, but they all come from the same raw dataset. The Real Time Scoring System is a dataset generated by the NHL that records select on ice actions. They record the time, the players on ice, the geographical location, etc. Using this data, all xGAR, xG, RAPM-type numbers are generated. This is know as a public model.

There private firms that generate their own datasets, but those are very expensive and usually only sell their data to professional sports team.

In terms of the actual terms used (Corsi, Fenwich, on-ice, etc.) this is a great starting point: Evolving-Hockey

At the end of this they link an article written by Charlie that explains more of the analytical side of things.

Awesome. Thank you!

 

[flyerslducks]

Provorov-jiricek paired up….could’ve been us.

Curious what others think but I wouldn’t have minded provorov on the flyers IIIFFFF he was put in his current cbj role. Not on any pp unit, 2nd pair (I think he’s getting pp time now, but he wasn’t when werenski and severson were healthy). Never would’ve happened and we never would’ve seen sanheim in his current situation if he hadn’t left but I blame the organization a lot for this; they should’ve given sanheim more opportunity early. I can’t remember the last time we developed a true number 1 d and I’m scared we will f*** it up again when we get the chance to draft one. We mishandled sanheim and provorov badly

Sanheim-jiricek
Provorov-york

Solid top 4 tbh

 

[Philly Fanatik]

Provorov-jiricek paired up….could’ve been us.

Curious what others think but I wouldn’t have minded provorov on the flyers IIIFFFF he was put in his current cbj role. Not on any pp unit, 2nd pair (I think he’s getting pp time now, but he wasn’t when werenski and severson were healthy). Never would’ve happened and we never would’ve seen sanheim in his current situation if he hadn’t left but I blame the organization a lot for this; they should’ve given sanheim more opportunity early. I can’t remember the last time we developed a true number 1 d and I’m scared we will f*** it up again when we get the chance to draft one. We mishandled sanheim and provorov badly

Sanheim-jiricek
Provorov-york

Solid top 4 tbh

I’m still not over missing on Makar or Heiskanan

 

[dats81]

Will he even play?

Depends on if/when Morgan Frost finds himself back in Torts's dog house.

 

[Striiker]

Any lineup that includes Provorov is a “HELL NO” from me

 

[Captain Dave Poulin]

Any lineup that includes Provorov is a “HELL NO” from me

I am PROUD to say "Hell no" to Provorov.

 

[Striiker]

I am PROUD to say "Hell no" to Provorov.

Shame on everyone who waited that long to become a Provorov hater

Should’ve been hating for years

YEARS

 

[Flyerfan4life]

I’m still not over missing on Makar or Heiskanan

NP says hi...

 

[ponder719]

There are beginners guides that explain concepts like Corsi and xGF (Hockey Advanced Analytics: What Are They & Why Are They Important?) then there are dense mathematical explanations of individual metrics. What doesn’t exist to the best of my knowledge is the broad description of how and why you might prefer X stat for Y player. The closest you can come is going to an individual site and reading how their own metrics work — MoneyPuck.com -About and How it Works for example.

There would be some money in it for someone who could write an easily digestible book covering that topic. I'd love to read that one.

 

[Rebels57]

Provorov is a jerk off and I have a firm "no jerk off" policy.

 

[JojoTheWhale]

There would be some money in it for someone who could write an easily digestible book covering that topic. I'd love to read that one.

There roughly was one, but 2016 was a lifetime ago at this point: Hockey Abstract Presents... Stat Shot book by Rob Vollman

Vollman is the Kings’ Director of Analytics now, so it’s not going to be him again.

 

[ponder719]

There roughly was one, but 2016 was a lifetime ago at this point: Hockey Abstract Presents... Stat Shot book by Rob Vollman

Vollman is the Kings’ Director of Analytics now, so it’s not going to be him again.

Nice. I'll need to look into that. Thanks for the rec!

 

[Curufinwe]

We mishandled sanheim and provorov badly

If hey have swapped roles & ice time it would have been better for the Flyers, but it wouldn't have changed who they are as players.

 

[BrindamoursNose]

Provorov is a jerk off and I have a firm "no jerk off" policy.

I have a very loose no jerkoff policy.

You can be a jerkoff as long as you're good. Proper jerkoff-to-good ratio.

 

[bobbythebrain]

C'mon Danny boy. Make it happen!!

 

[Hollywood Cannon]

Once again, it's not a rebuild.

 

[freakydallas13]

The trades a rebuilding team will make depend on where they are in the standings, lmao.

 

[Flyerfan4life]

Once again, it's not a rebuild.

#rebuild complete..

LOL. what a dumpster fire Org.
win a few games... ok f*** it forget the rebuild

 

[wankstifier]

"Folks" is code for "you ungrateful f***s"

 

[Tripod]

Jeff Marek show talking Philly.

Risto to Leafs talked about because Leafs are looking for RHD and "snarl". That's why the rule out Walker instead.

Also mentioned Seeler.

Philly in the driver's seat with these guys and are in no rush because they have been playing well.

 

[Beef Invictus]

Once again, it's not a rebuild.

It never was. Because they never changed anyone or any of the guiding philosophies running the team. This has always been obvious.