Pittsburgh Sports Media Gibberish: Bombulie ‎to restore order @ the Trib(II)

[billybudd]

He actually said that either 1) The pens don't know what they're talking about or 2) Yohe made up the quote from "a number of sources"

I believe his statement is correct.

That's actually worse. If you're going to accuse a reporter of fabricating a story, you damn well better have something more substantial than "the opinion expressed by unnamed sources is wrong."

 

[Pens1566]

That's actually worse. If you're going to accuse a reporter of fabricating a story, you damn well better have something more substantial than "the opinion expressed by unnamed sources is wrong."

Oh I don't know. He was employed by the trib ...

Either way, his point still stands. Outside of alien body snatchers, one of the two scenrios Sam listed is/was happening.

 

[Jacob]

Toni something? If it's who I'm thinking of, that person needs to be institutionalized. Six months ago, she got about 20 of her goons to try and get him fired over referring to his wife and kids as "his" wife and "his" kids to give you some idea of how insane she is.

Trying to dismantle the patriarchy starting with Pittsburgh sports writers' social media accts, bless her heart.

 

[Coastal Kev]

Opinions similar to yours is why the Pens are in so much long term trouble to begin with. Analytics is a process of using available data to help determine what is the best decision to make. It knows no bounds and can be applied to ANYTHING you do or decide on. Just because your "brain" cannot figure out how to apply it to NHL hockey operations, doesn't mean that there aren't 100s (most likely 1000s) of other people out there who can. This entire subject has been developing specifically in the NHL over the last 10 years, but you still can't understand it's usefulness?

The only thing you are "calling out loudly" is stating how little you actually understand about the game.

Really?? Do you really believe your first statement? It's opinions similar to mine that are the reason for the Pens long term failure? I could address this one point in great detail but it is so devoid of any rational critical thinking, I will instead .

Analytics are BS. Can you show me where analytics have been used in hockey and it's results have been both unique and valuable?

Here is the real problem with analytics, you don't provide your "genius" formula all of the information available. Therefore no matter what the results you come up with, the result is biased based on the information you deemed important to use create your formula.

I have a brother that is computer scientist and he tells me story after story of the waste associated with his field. Why does this occur you may ask? Because the people that he reports to in an organization do not understand how people in his field design and implement their course of action. I'm sure the Carnegie Mellon guy is brilliant. He was so brilliant that he was able to sell his services to a professional team without bringing anything new to the table except his opinion and a shiny program that verifies it.

 

[billybudd]

Really?? Do you really believe your first statement? It's opinions similar to mine that are the reason for the Pens long term failure? I could address this one point in great detail but it is so devoid of any rational critical thinking, I will instead .

Analytics are BS. Can you show me where analytics have been used in hockey and it's results have been both unique and valuable?

Here is the real problem with analytics, you don't provide your "genius" formula all of the information available. Therefore no matter what the results you come up with, the result is biased based on the information you deemed important to use create your formula.

I have a brother that is computer scientist and he tells me story after story of the waste associated with his field. Why does this occur you may ask? Because the people that he reports to in an organization do not understand how people in his field design and implement their course of action. I'm sure the Carnegie Mellon guy is brilliant. He was so brilliant that he was able to sell his services to a professional team without bringing anything new to the table except his opinion and a shiny program that verifies it.

Good chance the Penguins have hired him for application design and development, not scouting. That's what happened with the guy who designed war-on-ice's defunct spiritual predecessor (forget what it was called). Oilers(?) hired him to create and maintain a similar, internal data sorting application that other teams would not have access to.

 

[Mischa]

Really?? Do you really believe your first statement? It's opinions similar to mine that are the reason for the Pens long term failure? I could address this one point in great detail but it is so devoid of any rational critical thinking, I will instead .

Analytics are BS. Can you show me where analytics have been used in hockey and it's results have been both unique and valuable?

Here is the real problem with analytics, you don't provide your "genius" formula all of the information available. Therefore no matter what the results you come up with, the result is biased based on the information you deemed important to use create your formula.

I have a brother that is computer scientist and he tells me story after story of the waste associated with his field. Why does this occur you may ask? Because the people that he reports to in an organization do not understand how people in his field design and implement their course of action. I'm sure the Carnegie Mellon guy is brilliant. He was so brilliant that he was able to sell his services to a professional team without bringing anything new to the table except his opinion and a shiny program that verifies it.

I don't think there's an advanced stats guy out there who says you can use the numbers with no context. You have to know exactly what situations they were collected from and which situations will be ignored. I don't see how it's a problem to know over the course of their career about how good at maintaining possession a player/team is. The entire point is just to address trends and keep them in mind while making decisions


Also A stats guy would be in mathematics/statistics rather than computer scientist. And looking at his qualifications, he is indeed in stats and computational finance, he isn't a programmer (though he likely has some of the related skills)

 

[Honour Over Glory]

Man, wtf is this...Days of Our Lives?

 

[Freeptop]

Really?? Do you really believe your first statement? It's opinions similar to mine that are the reason for the Pens long term failure? I could address this one point in great detail but it is so devoid of any rational critical thinking, I will instead .

Analytics are BS. Can you show me where analytics have been used and it's results have been both unique and valuable?

Here is the real problem with analytics, you don't give your "genius" formula all of the information available. Therefore no matter what the results you come up with, the result is biased based on the information you deemed important to use create your formula.

I have a brother that is computer scientist and he tells me story after story of the waste associated with his field. Why does this occur you may ask? Because the people that he reports to in an organization do not understand how people in his field design and implement their course of action. I'm sure the Carnegie Mellon guy is brilliant. He was so brilliant that he was able to sell his services to a professional team without bringing anything new to the table except his opinion and a shiny program that verifies it.

"Data" is not the plural of "anecdote." You've got one anecdote and you're extrapolating it to not just one, but two entire fields of science (both computer science and statistics).

Incidentally, I am a computer scientist (specifically, a software engineer with over 16 years of experience, to go along with my degree from Carnegie Mellon). The majority of my employers certainly have understood how people in my field design and implement our course of action. If they didn't, we would have failed. Miserably. I'm sorry your brother has not had a similar quality of employer as I've been able to enjoy for the majority of my career so far.

Statistics in hockey is fairly young. Even what we call "advanced metrics" is actually extremely basic stuff. But what they've been able to do is show that some statistics that are currently gathered to correlate strongly with an ability to win games. Note that this doesn't mean that those statistics cause the team to win, but that they are indicators that the team is doing something well.

Sometimes these statistics only reinforce what we can intuitively figure out. This is not a bad thing. For instance, the idea that teams sit back and defend more when they have a lead late in a game has been statistically shown to be accurate. This is what is commonly called "score effects" by the hockey analytics community. That's why most stats people focus in on 5v5 "close" statistics, since those are the circumstances where score effects aren't impacting the data.

Other cases, though, sometimes should make teams reconsider their tactics. Such as the statistical analysis that shows that dumping the puck in and chasing after it is more likely to result in a scoring chance against than for. If you think about it, it makes sense - you're giving the puck away to the other team and only hoping to get it back. But entirely too many teams rely on dumping it in and "establishing a strong forecheck".

None of this means that someone can simply rely on statistics for everything. Context is important. But what statistics does do is help you determine where you ought to be looking harder. This is because of two basic concepts from psychology: 1. Confirmation bias and 2. Recency bias. The first says that you tend to only remember the things that confirm your belief, while the second says that what you've observed recently sticks out in your mind more than overall history. Statistics are a way to combat biases and force you to re-evaluate things that your emotions can cloud. Emotions can cause coaches to overplay someone just because they like them, rather than because they're actually good at hockey, and so when that player does something good, that's what sticks out to them. Recency bias strongly effects GMs when they're handing out contracts to players who just had a career year, especially if it was in the playoffs.

Being able to use statistics as a sort of check on your emotions is something that the Pens have clearly been lacking in recent years.

 

[Coastal Kev]

I don't think there's an advanced stats guy out there who says you can use the numbers with no context. You have to know exactly what situations they were collected from and which situations will be ignored. I don't see how it's a problem to know over the course of their career about how good at maintaining possession a player/team is. The entire point is just to address trends and keep them in mind while making decisions


Also A stats guy would be in mathematics/statistics rather than computer scientist. And looking at his qualifications, he is indeed in stats and computational finance, he isn't a programmer (though he likely has some of the related skills)

I understand this completely. The reason I provided this as an example is that the analytical guys are reporting to people who have zero understanding of how their conclusions are formulated. Does statistics hold value, of course. But show me a team built entirely with using advance stats and let's see the results.

 

[Winger for Hire]

Why does it always have to be one or the other. Why can't we use stats in conjunction with scouting?

 

[Mr. T]

But show me a team built entirely with using advance stats and let's see the results.

Nice straw man argument.

 

[Honour Over Glory]

I'm not a huge fan of these analytic stat fads that are the big thing right now. It has his pros but also its cons and right now I feel like it's this huge thing that people feel like they should live and die by and it will have them make fools of themselves when it stops them from making a decision that would help the team because of "stats" rather than taking more real life situations into account.

But whatever.

Back to some dip **** reporter and his tweets.

 

[Honour Over Glory]

Why does it always have to be one or the other. Why can't we use stats in conjunction with scouting?

It should be every possible thing that is out there to create a group/committee that is the best for the team and its future.

 

[Mischa]

I understand this completely. The reason I provided this as an example is that the analytical guys are reporting to people who have zero understanding of how their conclusions are formulated. Does statistics hold value, of course. But show me a team built entirely with using advance stats and let's see the results.

No one is going to entirely just use advanced stats because it doesn't make sense. It's only part of the larger puzzle. It's a tool that can help quantify larger trends that you would often overlook when trying to go over tape personally.

The reason why computer scientists often have trouble explaining the use of their tools to their bosses is because their bosses are often business types. If he were reporting to morehouse I could absolutely see how it would be useless but (id assume) he's reporting to hockey people who understand the concepts of possession and whatever other situations the stats model.

I like how karamos described the role of stats guys. Basically the same as scouts. They're just another group of people giving you insight into smaller things that the main coaching and front office staff don't have time for

Edit: freetop did much better

 

[Fordy]

I understand this completely.

liar

 

[Mr. T]

liar

 

[Coastal Kev]

No one is going to entirely just use advanced stats because it doesn't make sense. It's only part of the larger puzzle. It's a tool that can help quantify larger trends that you would often overlook when trying to go over tape personally.

The reason why computer scientists often have trouble explaining the use of their tools to their bosses is because their bosses are often business types. If he were reporting to morehouse I could absolutely see how it would be useless but (id assume) he's reporting to hockey people who understand the concepts of possession and whatever other situations the stats model.

I like how karamos described the role of stats guys. Basically the same as scouts. They're just another group of people giving you insight into smaller things that the main coaching and front office staff don't have time for

Edit: freetop did much better

That is a very fair response. I'm still not sold how you use stats in hockey when there are so many outlying variables that will effect the numbers you are trying to find.

 

[Penguinator]

Pghkev, did you read post #983?

 

[Honour Over Glory]

That is a very fair response. I'm still not sold how you use stats in hockey when there are so many outlying variables that will effect the numbers you are trying to find.

I...

Your avatar is cute?

 

[Coastal Kev]

Pghkev, did you read post #983?

I did, an impressive response indeed. But in conclusion, the op speaks to statistics being used to counter bias.


I believe that stats in hockey tell very little without bias. You can use stats all you want, but it's entirely up to the bias of the person using those stats and placing a level of importance to each stat.

 

[Fordy]

i'm just so confused by what the point is supposed to be

are there no variables involved in scouting, or something? eyes are infallible? even that's a faulty premise, that all standards that used to judge a player must be 100% accurate and fully describe them with no hidden variables to be useful? nothing like that exists. what could the objection possibly be

 

[Penguinator]

I did, an impressive response indeed. But in conclusion, the op speaks to statistics being used to counter bias.


I believe that stats in hockey tell very little without bias. You can use stats all you want, but it's entirely up to the bias of the person using those stats and placing a level of importance to each stat.

...and that's why Ventura should be so useful, by presenting the data in a way that the org can make sense of it so their bias is filtered through analytical context.