Ongoing NHL thread - Part III

[Darkstone42]

Hornqvist notching a hat trick, and Phil Bourque dropping his nickname "Horny" on the radio (a Horny Hatty is what happened tonight, in case you needed questionable imagery), makes the point yet again that hockey players are completely terrible at nicknames.

Or awesome at it, as you preferences dictate.

 

[dash]

Flames are 1-7-1 without Kris Russell - It's funny how the advanced stats guys never bring that up.

 

[Darkstone42]

Flames are 1-7-1 without Kris Russell - It's funny how the advanced stats guys never bring that up.

I'm going to claim two statistical points:

1. Small sample size (9 games is insufficient to make a definitive statement on value, as such a selection is far too subject to variance.)
2. Quality of replacement, which, if terrible, skews this argument in Russell's favor, but not by merit of his goodness so much as by merit of his replacement's badness.

He certainly may carry more value than the metrics capture, but that doesn't necessarily mean he's an objectively "good" defenseman.

 

[dash]

I was being quite facetious with my above post DS, I know there are many factors at play here, I guess I have trouble understanding the advocates of advanced stats referring to GMs who have played a major role in Stanley Cup championships teams as dummies when they acquire players whose advanced stats leave something to be desired.

/I guess I wonder at times if advanced stats leave something to be desired.

 

[DragonfromTO]

Flames are 1-7-1 without Kris Russell - It's funny how the advanced stats guys never bring that up.

9 freaking games dash. Come on man.

 

[Darkstone42]

I was being quite facetious with my above post DS, I know there are many factors at play here, I guess I have trouble understanding the advocates of advanced stats referring to GMs who have played a major role in Stanley Cup championships teams as dummies when they acquire players whose advanced stats leave something to be desired.

/I guess I wonder at times if advanced stats leave something to be desired.

Well, the factor exists that these statistics (and other, similar internal value systems) are only now gaining prevalence, and in an environment in which no one is optimizing their teams this way, someone still has to win. In those environments, whoever builds the best team not using those analytics has still built the best team.

Even so, whether intentional or not, most of the best teams in recent memory have been built for possession, which the advanced metric reward. Even thinking back to those Red Wings teams which always seemed to go deep in the playoffs, for example, they excelled at possessing the puck, both in terms of generating sustained pressure in the offensive zone and being able to outlet the puck from their own end.

There is, after all, an observation which led to the construction of these advanced stat models.

 

[dash]

Don't get me wrong, I know that teams that possess the puck usually end up being successful in the long run and I'm not trying to say that advanced stats are not useful (I truly believe they are). It is obviously much easier to apply these type of stats to the game of baseball than a fluid game like hockey though and I think there is still additional work and analysis that needs to be done.

 

[Darkstone42]

Don't get me wrong, I know that teams that possess the puck usually end up being successful in the long run and I'm not trying to say that advanced stats are not useful (I truly believe they are). It is obviously much easier to apply these type of stats to the game of baseball that a fluid game like hockey though and I think there is still additional work and analysis that needs to be done.

There are a few areas which could be used to improve the stats to be more predictive of outcomes. I think things like possession time in the offensive zone (clocks for each team measuring time the puck spends on and between sticks of offensive players in the attacking zone), broken down for when specific players are on the ice; a shot attempt/shot on goal for and against tally which scales with a gradient for distance from the net; a scoring chance version of possession stats; a stat which looks at second chance shots for and against; and various other ways to record not just how often a team has the puck and shoots at the other goalie, but how much offense they're generating when they do that (and vice versa on defense) are probably the direction this thing is heading, and many teams committed to analytics probably already have some at least rudimentary version of these models.

I'd guess Russell would score better under some of these due to his particular skill set, though I wonder by how much.

 

[Darkstone42]

Baseball has PitchF/X. Does hockey have ShotF/X? That would be great.

 

[DragonfromTO]

Baseball has PitchF/X. Does hockey have ShotF/X? That would be great.

The real breakthrough is going to be the data from SportVU. Seems like it would be much more valuable.

 

[dash]

A big +2 Matthews points for the Flames tonight - So this is what it feels like to be an Oilers fan? Man, in a weird way, you have to give Oilers fans credit for sticking it out for a decade of bad hockey.

 

[DragonfromTO]

I was being quite facetious with my above post DS, I know there are many factors at play here, I guess I have trouble understanding the advocates of advanced stats referring to GMs who have played a major role in Stanley Cup championships teams as dummies when they acquire players whose advanced stats leave something to be desired.

/I guess I wonder at times if advanced stats leave something to be desired.

Tons of professional baseball players wear titanium necklaces and have had great success while wearing them on the field. Despite having never played professional baseball myself I still feel OK questioning whether those necklaces have any effect at all, even though All-Stars and Hall of Famers who know way more about what it takes to be an MLB player than I do swear by them. Smart people are still wrong sometimes, even about the subjects that they are experts in.

 

[dash]

Smart people are still wrong sometimes, even about the subjects that they are experts in.

I can't argue with that statement.

/I mean, I was wrong about Brandon Bollig.

 

[DragonfromTO]

I can't argue with that statement.

/I mean, I was wrong about Brandon Bollig.

I don't know that you ever claimed expertise about the lower simians though dash

 

[dash]

I don't know that you ever claimed expertise about the lower simians though dash

Hey, they're my kind of people

 

[DragonfromTO]

Hey, they're my kind of people

They always know where all of the good fruit is

 

[Bloody Brian Burke]

NHL Lottery Simulator

I did it 20 times and the Leafs came in the top 3 only six times. Leafs may finish dead last and I'll be convinced they're picking 4th.

 

[jstewismybastardson]

NHL Lottery Simulator

I did it 20 times and the Leafs came in the top 3 only six times. Leafs may finish dead last and I'll be convinced they're picking 4th.

In the cesspool, theyve been doing it about 3 times each morning on air for the past 2 weeks and canucks have won once lol

 

[Bloody Brian Burke]

In the cesspool, theyve been doing it about 3 times each morning on air for the past 2 weeks and canucks have won once lol

I don't mind if they don't win it. TBH, there would be something noble about the worst team in the league not winning the lottery three years in a row (unless Edmonton gets it again). But fuck at least the second or third pick would be nice. Even that is a longshot now.

 

[forty_three]

Hornqvist notching a hat trick, and Phil Bourque dropping his nickname "Horny" on the radio (a Horny Hatty is what happened tonight, in case you needed questionable imagery), makes the point yet again that hockey players are completely terrible at nicknames.

Or awesome at it, as you preferences dictate.

"I will shorten your last name and add a 'Y' or an 'er'. AAAAND that's what we will call you from now until forever."

Flames are 1-7-1 without Kris Russell - It's funny how the advanced stats guys never bring that up.

INTANGIBLES

I was being quite facetious with my above post DS, I know there are many factors at play here, I guess I have trouble understanding the advocates of advanced stats referring to GMs who have played a major role in Stanley Cup championships teams as dummies when they acquire players whose advanced stats leave something to be desired.

/I guess I wonder at times if advanced stats leave something to be desired.

They have their places, just like a coaches gut has a place. Neither one should be trusted explicitly.

I would love for some advanced stats guru to map out the 1980 USA Olympic team's advanced stats. Or the 4 cups in a row Isles. Didn't feel like either of those teams was very possesionish. They just whittled away at you and pounced when you fucked up.