Arguments Against Advanced Stats?

[OccamsRazor]

Numbers help determine who wins, would you disagree with that?

 

No.

 

But other numbers i see many times over and over again i've seen the Flyers out shoot the other team by 20 or 30 shots and still lose so that is a perfect example of why shots don't tell the whole story.

 

And as far as Myers goes on a good Sabres team as a 18 year old kid he played 82 games scored 11 goals and 37 assists and was +13. The following year he score 10 goals. So i see him do it once i think he can get back there with the right coaching.

[RiskyBryzness]

Have you found that to be true?  I guess what I'm asking is have you found them to be generally predictive?   Would they have predicted the Kings beating the Hawks last year, for example?   I honestly don't know, but am asking since you seem to be fluent in it.

 

Or is it more "they would have predicted probably the last 4-8 teams playing but ultimately it can still come down to a bounce?"

The Blackhawks and Kings are two of the best possession teams in the NHL. Last season (I read somewhere, i'll try to find the source) the final 4 or so teams all were in the top 10 (or so) teams in puck possession (corsi).

[RiskyBryzness]

No.

 

But other numbers i see many times over and over again i've seen the Flyers out shoot the other team by 20 or 30 shots and still lose so that is a perfect example of why shots don't tell the whole story.

 

And as far as Myers goes on a good Sabres team as a 18 year old kid he played 82 games scored 11 goals and 37 assists and was +13. The following year he score 10 goals. So i see him do it once i think he can get back there with the right coaching.

I'm going to ignore the Myers discussion in this thread mostly because i'm trying to genuinely discuss adv stats here, and because I can't see Myers coming here.

 

And on that first comment, that happens sometimes, but there are also nearly just as many games where they are outshot and win, yesterday for example. However, it usually all balances out.

[OccamsRazor]

Numbers help determine who wins, would you disagree with that?

 

Look no further than last win versus LA.

 

Flyers had 16 total shots to LA's 38....hows that for corsi.

[doom88]

The Blackhawks and Kings are two of the best possession teams in the NHL. Last season (I read somewhere, i'll try to find the source) the final 4 or so teams all were in the top 10 (or so) teams in puck possession (corsi).

Correlation is not causation.

[ruxpin]

The Blackhawks and Kings are two of the best possession teams in the NHL. Last season (I read somewhere, i'll try to find the source) the final 4 or so teams all were in the top 10 (or so) teams in puck possession (corsi).

 

But the final four weren't necessarily the top four in corsi, so there is still a certain amount of bounce, yeah?  I mean, it's still the "you have to be good to be lucky" or "you have to put yourself in the position to take advantage of luck" but it's still not necessarily a 1 to 1, 2 to 2 result.  Correct? 

 

Sorry for the questions born of ignorance, but I find it an interesting topic.

[RiskyBryzness]

Look no further than last win versus LA.

 

Flyers had 16 total shots to LA's 38....hows that for corsi.

I've said multiple times within this thread and others that looking at it game by game doesnt tell the whole story, but over the course of a season it all balances out. See my post about LA/CHI a few replies up.

[RiskyBryzness]

Correlation is not causation.

I have not once mentioned causation.

[OccamsRazor]

I've said multiple times within this thread and others that looking at it game by game doesnt tell the whole story, but over the course of a season it all balances out. See my post about LA/CHI a few replies up.

 

 

Well we'll just agree to disagree.

[RiskyBryzness]

But the final four weren't necessarily the top four in corsi, so there is still a certain amount of bounce, yeah?  I mean, it's still the "you have to be good to be lucky" or "you have to put yourself in the position to take advantage of luck" but it's still not necessarily a 1 to 1, 2 to 2 result.  Correct? 

 

Sorry for the questions born of ignorance, but I find it an interesting topic.

No worries, I'm enjoying this discussion.

 

Theres clearly an amount of bounce as there is with any stat. One of the problems with Corsi is that it doesn't factor in goaltending, something that the Rangers had carry them last season.

 

Corsi correlates with success IMO, but there is not a causation.

[ruxpin]

I'm going to ignore the Myers discussion in this thread mostly because i'm trying to genuinely discuss adv stats here, and because I can't see Myers coming here.

 

And on that first comment, that happens sometimes, but there are also nearly just as many games where they are outshot and win, yesterday for example. However, it usually all balances out.

 

 

Question #67 born of ignorance:  So maybe the advanced stats are a better macro indicator than maybe just a single game?  It seems like the majority of the time they're probably indicative of a given game's result but enough that maybe it's better to use them in a 10-game or season-to-date sense.

[OccamsRazor]

I've said multiple times within this thread and others that looking at it game by game doesnt tell the whole story, but over the course of a season it all balances out.

 

And i'm not buying that argument sounds like you want to use it when it suits you.

[RiskyBryzness]

But other numbers i see many times over and over again i've seen the Flyers out shoot the other team by 20 or 30 shots and still lose so that is a perfect example of why shots don't tell the whole story.

 

 

Look no further than last win versus LA.

 

Flyers had 16 total shots to LA's 38....hows that for corsi.

 

 

Well we'll just agree to disagree.

 

Regression over the course of a period of games.

[RiskyBryzness]

Question #67 born of ignorance:  So maybe the advanced stats are a better macro indicator than maybe just a single game?  It seems like the majority of the time they're probably indicative of a given game's result but enough that maybe it's better to use them in a 10-game or season-to-date sense.

 

Yes exactly. Looking at a game has more flaws in the stats simply due to the smaller sample size. Looking at stats over a period of time, say 10-20 games or even longer, the stats even out to what is expected.

 

For example, in baseball if a player goes 4-for-4 with a home run, you can't expect that every game. Same thing here with advanced stats.

And i'm not buying that argument sounds like you want to use it when it suits you.

 

I use it over a period of time. I'm surprised you really haven't changed your stance at all. Looking at it game-by-game is more apt to flaws than looking at it over a larger sample size, just as it is with everything.

[doom88]

I have not once mentioned causation.

You are implicitly by pushing the questionable advanced stats so hard, and dismissing any resistance as beneath you.

[ruxpin]

No worries, I'm enjoying this discussion.

 

Theres clearly an amount of bounce as there is with any stat. One of the problems with Corsi is that it doesn't factor in goaltending, something that the Rangers had carry them last season.

 

Corsi correlates with success IMO, but there is not a causation.

 

That's fair.  So we've probably established that corsi, etc., gives a pretty good picture of a team--at least on a macro level.  Are you as confident with it--macro or micro--on the individual level?  I guess what I'm asking is do YOU have instances where you have been negative about a player having watched him or based on traditional stats and had a WTF moment when you've looked at the advanced stats?

[doom88]

And i'm not buying that argument sounds like you want to use it when it suits you.

Stop making sense, we work better when we disagree. Coburn is an NHL defenseman! Manning sucks!

[OccamsRazor]

Yes exactly. Looking at a game has more flaws in the stats simply due to the smaller sample size. Looking at stats over a period of time, say 10-20 games or even longer, the stats even out to what is expected.

 

For example, in baseball if a player goes 4-for-4 with a home run, you can't expect that every game. Same thing here with advanced stats.

 

I use it over a period of time. I'm surprised you really haven't changed your stance at all. Looking at it game-by-game is more apt to flaws than looking at it over a larger sample size, just as it is with everything.

 

There really isn't a stance to change really it's numbers like +/- i look at them but the same i watch some of these guys play and it doesn't justify it. If you like the corsi stats thats fine i don't i go more on what i see on my TV.

[OccamsRazor]

Stop making sense, we work better when we disagree. Coburn is an NHL defenseman! Manning sucks!

 

 

Coburn sucks too then. I agree. 

[RiskyBryzness]

You are implicitly by pushing the questionable advanced stats so hard, and dismissing any resistance as beneath you.

 

These stats aren't questionable. They've been proven to correlate with success. Obviously there is some wiggle room as there is with anything.

 

That's fair.  So we've probably established that corsi, etc., gives a pretty good picture of a team--at least on a macro level.  Are you as confident with it--macro or micro--on the individual level?  I guess what I'm asking is do YOU have instances where you have been negative about a player having watched him or based on traditional stats and had a WTF moment when you've looked at the advanced stats?

I'm glad you're at least getting something out of this lol.

 

I typically watch a game and don't look at the advanced stats till after a period or game. I'm surprised a decent amount at first, but then looking/thinking back I understand it more.

 

As far as macro/micro, I have a similar level of confidence with it individually as I do with it as a team. A player's usage and other things can change though, just like a team's system, coach and personnel can change.

[OccamsRazor]

Coburn is an NHL defenseman! Manning sucks!

 

Like for example with +/- it won't show..............Coburn on the ice for 10 PP goals against the last 13 games oooooooch......please spin that as a positive please!

[RiskyBryzness]

Stop making sense, we work better when we disagree. Coburn is an NHL defenseman! Manning sucks!

 

Haha I actually like Manning and think he should be up, but not for Coburn.

 

There really isn't a stance to change really it's numbers like +/- i look at them but the same i watch some of these guys play and it doesn't justify it. If you like the corsi stats thats fine i don't i go more on what i see on my TV.

Agree to disagree then. Some people like different stats, but its just nice to have some flexibility in your viewpoints.

 

 

To both of you, you can look forward to a breakdown of Coburn similar to the one about Grossmann in the coming days/weeks haha.

[RiskyBryzness]

Like for example with +/- it won't show..............Coburn on the ice for 10 PP goals against the last 13 games oooooooch......please spin that as a positive please!

Coburn has been **** on the PK recently, theres no arguing against that. He has made up for it with his play at 5v5 though IMO.

[RiskyBryzness]

I look forward to continuing this later, but i'm out to run a few errands now then watch the birds!

[doom88]

Like for example with +/- it won't show..............Coburn on the ice for 10 PP goals against the last 13 games oooooooch......please spin that as a positive please!

Here you go!