The Advanced Stats Thread Episode V: Rick Nash Camera Stares/60

[silverfish]

Oh please...

Care to elaborate?

 

[ReggieDunlop68]

Care to elaborate?

Can you?

You took one quick stat that simply said no attempt is taken within 10 seconds , to gloss over debate about zone starts.

 

[silverfish]

Can you?

You took one quick stat that simply said no attempt is taken within 10 seconds , to gloss over debate about zone starts.

Sure, that's what that post is. But, this is not one metric to gloss over debate about zone starts, it's another metric. Very different.

Believe me, I've done extensive research on zone starts. They matter, but they don't matter nearly as much as you'd believe they should.

If you have a dissenting argument, I'd be happy to hear it.

 

[Mac n Gs]

Zone starts though

Are there studies with more time points for this analysis? What made them choose 10 seconds and why is that significant?

Edit: I'm asking because 10 seconds seems kinda arbitrary and would lead more to a team forcing a low-percentage shot off the faceoff.

 

[silverfish]

Are there studies with more time points for this analysis? What made them choose 10 seconds and why is that significant?

I'd imagine Dellow chose 10 seconds because that is the adjustment that's used on http://stats.hockeyanalysis.com/. (Don't quote me on that, just an assumption).

Zone Start Adjusted statistics remove the 10 seconds immediately after a zone face off. It has been shown that the majority of the benefit or penalty of a zone start occurs during the first 10 seconds.

https://puckalytics.com/#/glossary

 

[ReggieDunlop68]

Sure, that's what that post is. But, this is not one metric to gloss over debate about zone starts, it's another metric. Very different.

Believe me, I've done extensive research on zone starts. They matter, but they don't matter nearly as much as you'd believe they should.

If you have a dissenting argument, I'd be happy to hear it.

Never mind.

I assumed it was a "zone starts smh...". I'm not saying they are the be and and end all either.

 

[silverfish]

Never mind.

I assumed it was a "zone starts smh...". I'm not saying they are the be and and end all either.

Well, saying anything but "zone starts smh..." is kind of difficult not to do. Because there really is a lot of evidence that suggests that they are not all that important. And it's hard to have a discussion about zone starts, because the other side of the argument essentially boils down to:

"Of course they matter! They have to!"

Without any reasonable objective evidence that shows the actual impact of zone starts on a player's shot attempt metrics.

 

[Mac n Gs]

I'd imagine Dellow chose 10 seconds because that is the adjustment that's used on http://stats.hockeyanalysis.com/. (Don't quote me on that, just an assumption).



https://puckalytics.com/#/glossary

Ah gotcha, although I'd be interested in seeing the data that went into making that claim. That's A LOT of minutiae to track (unless it's just being scrapped from nhl play-by-play)

 

[silverfish]

Ah gotcha, although I'd be interested in seeing the data that went into making that claim. That's A LOT of minutiae to track (unless it's just being scrapped from nhl play-by-play)

Definitely working off of scraped NHL pbp data.

Manny actually has all the PBP data from 07-08 through 15-16 here in R Data files... If you're computer has the RAM to handle it, could do some solid analysis with all this...

http://www.corsica.hockey/data/

 

[ReggieDunlop68]

Well, saying anything but "zone starts smh..." is kind of difficult not to do. Because there really is a lot of evidence that suggests that they are not all that important. And it's hard to have a discussion about zone starts, because the other side of the argument essentially boils down to:

"Of course they matter! They have to!"

Without any reasonable objective evidence that shows the actual impact of zone starts on a player's shot attempt metrics.

First, if the only way to say an offensive zone start is a benefit is by seeing how much it effects an individuals CF%, then yeah it's easy to draw a fall negative.

For example, instead of a single zone start, I would look at the cumulative effect of zone starts.

One could also look at how long until an attempt is taken by the team with the offensive zone face off even if it leaves the zone, but prior to another face off.

Just starting in the offensive zone creates more work for the opposing team since they must carry across the ice to finally have an attempt.

 

[silverfish]

First, if the only way to say an offensive zone start is a benefit is by seeing how much it effects an individuals CF%, then yeah it's easy to draw a fall negative.

For example, instead of a single zone start, I would look at the cumulative effect of zone starts.

One could also look at how long until an attempt is taken by the team with the offensive zone face off even if it leaves the zone, but prior to another face off.

Just starting in the offensive zone creates more work for the opposing team since they must carry across the ice to finally have an attempt.

I'm not entirely certain what you're getting at here :/

 

[ReggieDunlop68]

I'm not entirely certain what you're getting at here :/

Hockey is a fluid game and you guys try to make it to static.

 

[silverfish]

Hockey is a fluid game and you guys try to make it to static.

I don't agree.

Hockey is a fluid game, I'm just trying to understand the things I admit my eyes aren't good at catching. I'm not trying to "make it static".

Following the puck and 12 players at the same time 82+ games a season (I don't know about you, but I probably end up watching somewhere around 200 games a year), 60 minutes a game, is not the easiest thing in the world. Then to retain all that information? Nope. I can't do it. Bless anyone who can.

 

[Zil]

Analytics run baseball and they're taking over basketball. It's only a matter of time until hockey goes the same way. It's already started to some degree, although their influence isn't too noticeable in most front offices. There's a giant windfall to be had by whatever teams step to the forefront. Undervalued players are everywhere right now, just like they were in baseball in the late 90s and early 00s. It's not gonna stay that way forever.

 

[silverfish]

Analytics run baseball and they're taking over basketball. It's only a matter of time until hockey front offices go the same way. It's already started to some degree, although their influence isn't too noticeable in most front offices. There's a giant windfall to be had by whatever teams step to the forefront. Undervalued players are everywhere right now, just like they were in baseball in the late 90s and early 00s. It's not gonna stay that way forever.

Canes, Panthers, Leafs, Wild, Yotes. That's just off the top of my head, but no doubt there are others. Rangers have an "analytics guy", who knows what he does though

I am really excited to see how this old molds out and takes shape over the next 3-5 years. In a capped sport, you need to optimize your $ spent.

Also:

 

[Mac n Gs]

We do use analytics. MSG has Vally talking about shot quality and all that stuff all the time. AV just referenced possession time the other day. Chris Boyle works for MSG. We just don't have access to what they do.

 

[Bob Richards]

I don't agree.

Hockey is a fluid game, I'm just trying to understand the things I admit my eyes aren't good at catching. I'm not trying to "make it static".

Following the puck and 12 players at the same time 82+ games a season (I don't know about you, but I probably end up watching somewhere around 200 games a year), 60 minutes a game, is not the easiest thing in the world. Then to retain all that information? Nope. I can't do it. Bless anyone who can.

The cats part is unnecessary, but what they're getting at isn't.

 

[Zil]

We do use analytics. MSG has Vally talking about shot quality and all that stuff all the time. AV just referenced possession time the other day. Chris Boyle works for MSG. We just don't have access to what they do.

They use it. I just don't think they value it as highly as they should. I don't think it drives their personnel moves. It clearly doesn't rate high enough in the decision making process to stop them from heaping minutes onto guys like Girardi and Staal. They even called up Glass for no good reason.

 

[silverfish]

The cats part is unnecessary, but what they're getting at isn't.

Yep, exactly. I remember posting that a while back. I think Trxjw or nyr2k ended up beasting it, but I sucked

Eye-test is useful, but it's totally flawed. That's my take on hockey these days.

This is still my favorite, but it's been posted everywhere now, so most people know it.

 

[Bob Richards]

Yep, exactly. I remember posting that a while back. I think Trxjw or nyr2k ended up beasting it, but I sucked

Don't get me wrong, I don't doubt that a lot of our guys can carefully read these plays with fluency.

But when you starting piling shifts onto shifts onto games onto weeks onto months and seasons, things get lost.

Eye-test is useful, but it's totally flawed. That's my take on hockey these days.

Agreed.

I never understood the sentiment that hockey analysis is either one or the other with regards to the eye test/nerd numbers. Why can't we all meet in a healthy middle?

 

[2020 Cup Champions]

The cats part is unnecessary, but what they're getting at isn't.

I strongly suspect the cats part is necessary. I'm at a loss to explain it in detail, but I'd be surprised if they're not playing with the working memory and diffused learning.

 

[Bob Richards]

I strongly suspect the cats part is necessary. I'm at a loss to explain it in detail, but I'd be surprised if they're not playing with the working memory and diffused learning.

I know what you're getting at but I mean it in the most literal of senses. 2 minutes of cats.

 

[Irishguy42]

The eye test and advanced stats are both useful but flawed when used (and relied upon) separately. When used together, you'll have the best understanding of what is going on.

This needs to be the PSA every time and 'analytics vs eye test' discussion is brought up.

 

[Thirty One]

The key to being smart at hockey is not watching cat videos.

One of the big issues I have with the eye test is how it works alongside with confirmation bias. There are so many plays that happen during a game, what we often remember are 1) the big plays and b) the plays that confirm what we already thought. I wrote a story last year about 1) with Keith Yandle as the protagonist. And I'm not saying that b) is a problem with one side of this debate. I see lots of it on either side, and I'm guilty of it too. Damned near impossible not to be.

 

[Thirty One]

The eye test and advanced stats are both useful but flawed when used (and relied upon) separately. When used together, you'll have the best understanding of what is going on.

This needs to be the PSA every time and 'analytics vs eye test' discussion is brought up.

TheMoreYouKnow.gif