The Advanced Stats Thread Episode IX

[Machinehead]

Don’t even bring up +-. If you went by it, Shatty would be the 2nd best D on the team, and Stionk would be the worst player on the team by far.

Even meaningless stats aren’t in Pionk’s favor. Literally all Pionk has going for him is a few flashy offensive plays, he is egregiously bad defensively. I haven’t even followed the metrics closely this year, he just looks absolutely putrid out there when not showing some offensive flash.

In Pionk's case, do yourself a favor and don't. It's worse than you think.

 

[aufheben]

What does this board think of K rating from Corsica?

 

[Machinehead]

What does this board think of K rating from Corsica?

I think it has some value but there's a lot of bizarre shit that comes out of it.

For example, it says we're currently the 4th best team in the league. I don't know if this is the 4th best team in New York state.

 

[DanielBrassard]

I think it has some value but there's a lot of bizarre **** that comes out of it.

For example, it says we're currently the 4th best team in the league. I don't know if this is the 4th best team in New York state.

I would agree, I don't believe they are anywhere close to the 4th best team but there is definitely a big improvement in limiting quality chances. Maybe there is something to it.

 

[Machinehead]

I would agree, I don't believe they are anywhere close to the 4th best team but there is definitely a big improvement in limiting quality chances. Maybe there is something to it.

It's not AV hockey, I have to give Quinn that much.

 

[DanielBrassard]

It's not AV hockey, I have to give Quinn that much.

Heh true. AV hockey works when you have elite forward depth and 2 top defenseman (McDonagh and Stralman).

 

[SnowblindNYR]

I assume this has been done before and I'm not breaking new ground but I ran a regression, regular season PDO as an independent variable and playoff PDO as a dependent variable from 07 through 18 for all playoff teams. This is what I got:

	Coefficients	Standard Error	t Stat	P-value	Lower 95%	Upper 95%	Lower 95.0%	Upper 95.0%
Intercept	82.18609507	22.07932736	3.72230973	0.000266319	38.60831566	125.7638745	38.60831566	125.7638745
PDO	0.169548188	0.219849276	0.77120194	0.441633245	-0.264366446	0.603462822	-0.264366446	0.603462822

[TBODY] [/TBODY]

As you can see it's a poor predictor. Significance is a p-value less than 0.05. It's highly insignificant (not sure if that's a thing, but it's clearly not significant). So having a high PDO in the regular season doesn't necessarily predict a high PDO in the playoffs. So does that mean that this confirms that it's a measure of luck and that teams tend to not have high PDOs due to players or strategy? Not sure, but these are the results.

@Machinehead

 

[Machinehead]

I assume this has been done before and I'm not breaking new ground but I ran a regression, regular season PDO as an independent variable and playoff PDO as a dependent variable from 07 through 18 for all playoff teams. This is what I got:

	Coefficients	Standard Error	t Stat	P-value	Lower 95%	Upper 95%	Lower 95.0%	Upper 95.0%
Intercept	82.18609507	22.07932736	3.72230973	0.000266319	38.60831566	125.7638745	38.60831566	125.7638745
PDO	0.169548188	0.219849276	0.77120194	0.441633245	-0.264366446	0.603462822	-0.264366446	0.603462822

[TBODY] [/TBODY]

As you can see it's a poor predictor. Significance is a p-value less than 0.05. It's highly insignificant. So having a high PDO in the regular season doesn't necessarily predict a high PDO in the playoffs. So does that mean that this confirms that it's a measure of luck and that teams tend to not have high PDOs due to players or strategy? Not sure, but these are the results.

@Machinehead

This is very significant data. Nice work.

 

[SnowblindNYR]

This is very significant data. Nice work.

Thanks. I assume you're not talking about statistical significance, haha.

 

[silverfish]

 

[Machinehead]

Kreider had his eggs this morning.

 

[SnowblindNYR]

I think it has some value but there's a lot of bizarre **** that comes out of it.

For example, it says we're currently the 4th best team in the league. I don't know if this is the 4th best team in New York state.

Aren't you doing what you criticize eye test people of doing? Ignoring the stats that you don't like?

 

[Machinehead]

Aren't you doing what you criticize eye test people of doing? Ignoring the stats that you don't like?

I guess.

Do you think there's any validity to the idea that the Rangers are a top 5 team?

 

[SnowblindNYR]

I guess.

Do you think there's any validity to the idea that the Rangers are a top 5 team?

No, but I also don't think they're as bad as you make them out to be. How is K calculated anyway?

 

[Mac n Gs]

No, but I also don't think they're as bad as you make them out to be. How is K calculated anyway?

A Brief Introduction To K | Corsica

 

[SnowblindNYR]

A Brief Introduction To K | Corsica

Thanks

 

[SnowblindNYR]

A Brief Introduction To K | Corsica

Interesting, I learned what binomial logistical regression means but don't remember now. Regression can get complicated.

 

[Synergy27]

Interesting read/opinion piece on potential misuse of predictive modeling:

East Coast Offense: Own Your Mistakes

Chris Liss is my favorite "fantasy sports" analyst. He is moreso a bettor/handicapper and seems legitimately knowledgable in statistics and philosophy. Curious to see what some of you die hard stats guys think about this.

 

[Machinehead]

Again, we're not doing the corsi but winning the expected goals tonight.

Is anyone else ok with this right now, but would like to be better in shot metrics when we have a more competitive team?

 

[Irishguy42]

Again, we're not doing the corsi but winning the expected goals tonight.

Is anyone else ok with this right now, but would like to be better in shot metrics when we have a more competitive team?

Well here's a better question: How much of the poor shot metrics/good xG measurements on the season is being driven by the lack of talent on the backend/lack of star talent on the front end?

There is something to be said for instilling a system first. I don't think this year would have looked that good anyway, especially with a rookie coach. I would look more at how the team responds next season after this "transition" year from AV. But again, the talent will change, so...

 

[Machinehead]

Well here's a better question: How much of the poor shot metrics/good xG measurements on the seasons is being driven by the lack of talent on the backend/lack of star talent on the front end?

There is something to be said for instilling a system first. I don't think this year would have looked that good anyway, especially with a rookie coach. I would look more at how the team responds after this "transition" year from AV.

I tend to agree with this.

I think Quinn has done a fine job with a flawed roster and it should take care of itself when the roster gets better.

Plus, it's not wonderful hockey, but it's not eye-gouging hockey. I can digest this. Which is good, because rebuild or no rebuild, I still have to watch 82 games of it.

 

[Irishguy42]

I tend to agree with this.

I think Quinn has done a fine job with a flawed roster and it should take care of itself when the roster gets better.

Plus, it's not wonderful hockey, but it's not eye-gouging hockey. I can digest this. Which is good, because rebuild or no rebuild, I still have to watch 82 games of it.

Yes. Big.

 

[silverfish]

 

[DanielBrassard]

Well here's a better question: How much of the poor shot metrics/good xG measurements on the season is being driven by the lack of talent on the backend/lack of star talent on the front end?

There is something to be said for instilling a system first. I don't think this year would have looked that good anyway, especially with a rookie coach. I would look more at how the team responds next season after this "transition" year from AV. But again, the talent will change, so...

I'm sure that's part of it, but then how do you explain teams like Montreal and New Jersey who on paper don't have better talent, outside of Hall? Montreal is one of the better CF% teams in the league and the Devils are more than respectable. I don't understand why the Rangers are just so bad. And yes, they have been good at limiting quality shots against, and good at getting quality shots for, but that remains to be seen how long that lasts. Only team that does it consistently while being outshot consistently is Minnesota.

 

[Mac n Gs]

Again, we're not doing the corsi but winning the expected goals tonight.

Is anyone else ok with this right now, but would like to be better in shot metrics when we have a more competitive team?

You fought with me last year about this re: Minnesota