All Purpose Analytics and Extended Stats Discussion

[ChibiPooky]

It's not my opinion that you need to limit variables and have "controls" in experimental situations. That's how science and data work. If you have faulty methodology that doesn't account for variables then the interpretation of the resulting data will always need to be couched with caveats about its accuracy. There are many who don't want to do that, and that's what I mean with the hammer analogy. It may be better than the stick we were using before but it's still not right for everything we're trying to do with it.

It's how science works, but it's not how real-world data works. Real-world data is dirty as hell. The whole couching with caveats bit is a lot more about the data than the analysis in this case. And because of the number of events we're working with, cleaning the data takes away a lot of value.

And just because it's not perfect doesn't mean you don't use it. You use it while understanding how it's not perfect while looking for something better. And if you don't have a stat guy, you're probably neither using the "stick" to its full potential (or maybe even correctly) nor looking for a better one.

Do you really think these stat guys are going into meetings saying "this guy has 45% Corsi so he has to go" or the like? Your whole argument seems to be motivated by the idea that stat guys are basically message board fans with influence. They're not.

 

[g00n]

It's how science works, but it's not how real-world data works. Real-world data is dirty as hell. The whole couching with caveats bit is a lot more about the data than the analysis in this case. And because of the number of events we're working with, cleaning the data takes away a lot of value.

And just because it's not perfect doesn't mean you don't use it. You use it while understanding how it's not perfect while looking for something better. And if you don't have a stat guy, you're probably neither using the "stick" to its full potential (or maybe even correctly) nor looking for a better one.

Do you really think these stat guys are going into meetings saying "this guy has 45% Corsi so he has to go" or the like? Your whole argument seems to be motivated by the idea that stat guys are basically message board fans with influence. They're not.

None of this changes what I said...use these tools for what they actually show, do not abuse them and make them into something they're not.

I don't know what goes on in the player personnel meetings and I don't think you do, either. But I do know what fans say (massive abuse, usually) and I know the industry is using these stats. I don't need to extrapolate it to apply equally to everyone, some kind of usefulness is implied. If all these teams are hiring fancy stat gurus just to stand around and shrug their shoulders, then why are they bothering? They're making some kind of determination from the data, and if it isn't based on these same kinds of stat comparisons then what is it? Do you know?

 

[Burakovsky95*]

No bother arguing with people still stuck in the Gretzky era of stats

 

[ChibiPooky]

But I do know what fans say (massive abuse, usually) and I know the industry is using these stats. I don't need to extrapolate it to apply equally to everyone, some kind of usefulness is implied.

If you don't see a difference between fans and professional statisticians, then there is literally no point in this conversation.

Let's put this in context: According to the New York Times, "Adulthood does not alleviate our quantitative deficiency. A 2012 study comparing 16-to-65-year-olds in 20 countries found that Americans rank in the bottom five in numeracy. On a scale of 1 to 5, 29 percent of them scored at Level 1 or below, meaning they could do basic arithmetic but not computations requiring two or more steps."

Americans are bad at math to the level that they can't even begin to understand what these stats are. That's the "fan" population we're talking about here. There's a reason these people seem to be "hitting everything with the same hammer." It's not a hammer, but that's what the majority of people see it as.

 

[g00n]

And again, I'm not opposed to professionals applying the best resources available to the craft. I'm against the abuse of that data and the use of it to the exclusion of all context or observational analysis.

Fans who have recently discovered these stats do this regularly when they lean too heavily on the raw numbers. Or when they look at a 51% vs a 49% and determine the first guy is clearly superior. It happens all the time and I'm pretty sure that's a big part of why this thread exists...

 

[g00n]

If you don't see a difference between fans and professional statisticians, then there is literally no point in this conversation.

See my previous post. You're knocking down a strawman.

 

[g00n]

No bother arguing with people still stuck in the Gretzky era of stats

So you ask what the arguments are against the way you abuse statistics, and when provided with a wealth of counter-arguments that include supporting quotes from the founder of the stats you stick your fingers in your ears.


edit: only talked about ignore list because this kid has repeatedly responded to myself and others by telling them to block him

 

[ChibiPooky]

And again, I'm not opposed to professionals applying the best resources available to the craft. I'm against the abuse of that data and the use of it to the exclusion of all context or observational analysis.

Fans who have recently discovered these stats do this regularly when they lean too heavily on the raw numbers. Or when they look at a 51% vs a 49% and determine the first guy is clearly superior. It happens all the time and I'm pretty sure that's a big part of why this thread exists...

Right. I'm saying that professionals are unlikely to abuse the data and use it to the exclusion of context or analysis.

 

[Burakovsky95*]

If you don't see a difference between fans and professional statisticians, then there is literally no point in this conversation.

Let's put this in context: According to the New York Times, "Adulthood does not alleviate our quantitative deficiency. A 2012 study comparing 16-to-65-year-olds in 20 countries found that Americans rank in the bottom five in numeracy. On a scale of 1 to 5, 29 percent of them scored at Level 1 or below, meaning they could do basic arithmetic but not computations requiring two or more steps."

Americans are bad at math to the level that they can't even begin to understand what these stats are. That's the "fan" population we're talking about here. There's a reason these people seem to be "hitting everything with the same hammer." It's not a hammer, but that's what the majority of people see it as.

Isn't everything Corsi related extremely easy to calculate? Except for maybe Corsi Rel QoC but even that is still simple math.

 

[g00n]

Right. I'm saying that professionals are unlikely to abuse the data and use it to the exclusion of context or analysis.

Let's hope so, but....Do you know this, or are we assuming? Can anyone show exactly how these stats are being used by the professionals? Are we going to assume all the teams in the league are equally equipped in this regard?

Because whenever this topic comes up on discussion boards the fans defend their abuses of the stats by pointing to the teams that are using the stats. Now we're saying they aren't the same applications. Then what are the applications?

 

[g00n]

Isn't everything Corsi related extremely easy to calculate? Except for maybe Corsi Rel QoC but even that is still simple math.

It's not the difficulty of calculation, it's the comprehension of how a simple to calculate stat is not itself comprehensive.

 

[ChibiPooky]

Let's hope so, but....Do you know this, or are we assuming? Can anyone show exactly how these stats are being used by the professionals? Are we going to assume all the teams in the league are equally equipped in this regard?

Because whenever this topic comes up on discussion boards the fans defend their abuses of the stats by pointing to the teams that are using the stats. Now we're saying they aren't the same applications. Then what are the applications?

I would categorize this also under "fan abuse of data" rather than a potential indication that the pros aren't doing things right. We don't know how these teams use the data or even what data they collect. The applications are there, but it takes a ton of expertise to know specifically what they are, or as has been the case since hockey sabermetrics is a very new field, essentially invent them. We also don't have that expertise.

And no, I don't think the playing field is level among all teams. That's why I'm excited about this hire. It puts the Caps in a position to be near the top of the curve, rather than slipping to the bottom like I feared. The Caps still have their hockey guys, but every little thing you can add helps.

 

[ChibiPooky]

Isn't everything Corsi related extremely easy to calculate? Except for maybe Corsi Rel QoC but even that is still simple math.

Yes, the calculation itself is easy. That's part of the danger g00n is talking about, though. With a fairly simple end result, it's easy to fool yourself into thinking you have a better understanding of something than you actually do. Understanding what the statistic actually is and what it actually tells you and to what degree of certainty is significantly more complex.

 

[g00n]

I would categorize this also under "fan abuse of data" rather than a potential indication that the pros aren't doing things right. We don't know how these teams use the data or even what data they collect. The applications are there, but it takes a ton of expertise to know specifically what they are, or as has been the case since hockey sabermetrics is a very new field, essentially invent them. We also don't have that expertise.

And no, I don't think the playing field is level among all teams. That's why I'm excited about this hire. It puts the Caps in a position to be near the top of the curve, rather than slipping to the bottom like I feared. The Caps still have their hockey guys, but every little thing you can add helps.

And that's fine. But dispute over this topic is such that a thread debating it was necessary. We can't ignore fan abuse of stats in the very thread created by it.

So how would the moderators like to handle this topic in the future?

 

[Burakovsky95*]

Yes, the calculation itself is easy. That's part of the danger g00n is talking about, though. With a fairly simple end result, it's easy to fool yourself into thinking you have a better understanding of something than you actually do. Understanding what the statistic actually is and what it actually tells you and to what degree of certainty is significantly more complex.

Even that part isn't very hard with simple study. I understand that some fans (myself included sometimes) can be easily deceived in some instances but anyone with common sense can delve deeper into the numbers or context and realize what they're exactly looking at may not be the absolute finite truth. I'm in agreement with you about how these developed/seasoned hockey minds can sometimes trump all statistics but saying that the common fan is completely blind in these matters is off base. I don't think that's what you're exactly trying to say here but I feel like these stats have almost developed enough so that "danger" he's talking about really isn't that common if you can think at a high school level.

 

[ChibiPooky]

And that's fine. But dispute over this topic is such that a thread debating it was necessary. We can't ignore fan abuse of stats in the very thread created by it.

So how would the moderators like to handle this topic in the future?

It's not off topic, and it will be treated like any other discussion. If everyone follows the site rules, everything is peachy.

 

[g00n]

Even that part isn't very hard with simple study. I understand that some fans (myself included sometimes) can be easily deceived in some instances but anyone with common sense can delve deeper into the numbers or context and realize what they're exactly looking at may not be the absolute finite truth. I'm in agreement with you about how these developed/seasoned hockey minds can sometimes trump all statistics but saying that the common fan is completely blind in these matters is off base. I don't think that's what you're exactly trying to say here but I feel like these stats have almost developed enough so that "danger" he's talking about really isn't that common if you can think at a high school level.

In such cases, this might be in play: http://en.wikipedia.org/wiki/Dunning–Kruger_effect

 

[ChibiPooky]

Even that part isn't very hard with simple study. I understand that some fans (myself included sometimes) can be easily deceived in some instances but anyone with common sense can delve deeper into the numbers or context and realize what they're exactly looking at may not be the absolute finite truth. I'm in agreement with you about how these developed/seasoned hockey minds can sometimes trump all statistics but saying that the common fan is completely blind in these matters is off base. I don't think that's what you're exactly trying to say here but I feel like these stats have almost developed enough so that "danger" he's talking about really isn't that common if you can think at a high school level.

Put it this way: the professionals are saying that they have to be cautious because a) the field is so new, and b) the understanding of the field is still very limited even at its most advanced. There's a big danger in going further than has been proved to be valid or useful. The current stats have been fairly well established, but I know I haven't read everything about them - quite far from it in fact. And of course a big part of the danger can be thinking it's not there. As a guy who majored in math, I can safely say that it is a big long string of checking all your assumptions, all the way down through all your dependencies, and making sure literally everything you've done makes sense. There's way more to it than I can bring to the table though, especially when just tossing out a message board post.

 

[Burakovsky95*]

Put it this way: the professionals are saying that they have to be cautious because a) the field is so new, and b) the understanding of the field is still very limited even at its most advanced. There's a big danger in going further than has been proved to be valid or useful. The current stats have been fairly well established, but I know I haven't read everything about them - quite far from it in fact. And of course a big part of the danger can be thinking it's not there. As a guy who majored in math, I can safely say that it is a big long string of checking all your assumptions, all the way down through all your dependencies, and making sure literally everything you've done makes sense. There's way more to it than I can bring to the table though, especially when just tossing out a message board post.

I definitely agree, but I hate seeing people getting flamed for using the stats we currently have. I don't think anyone is putting their 100% faith in what we have now and they shouldn't but dismissing them altogether like many do on this board is just stupid IMO.

 

[ChibiPooky]

I definitely agree, but I hate seeing people getting flamed for using the stats we currently have. I don't think anyone is putting their 100% faith in what we have now and they shouldn't but dismissing them altogether like many do on this board is just stupid IMO.

There's definitely a balance to be struck.

If you think you see flaming, please use the report button at the bottom left of the post.

 

[NobodyBeatsTheWiz]

I definitely agree, but I hate seeing people getting flamed for using the stats we currently have. I don't think anyone is putting their 100% faith in what we have now and they shouldn't but dismissing them altogether like many do on this board is just stupid IMO.

Why?

Without proper context, which we don't have, these stats have no more utility than +/-.

Would you have a problem with people being flamed for using +/- on a game-to-game basis?

 

[Burakovsky95*]

Why?

Without proper context, which we don't have, these stats have no more utility than +/-.

Would you have a problem with people being flamed for using +/- on a game-to-game basis?

Explain to me how Zone adjusted relative corsi numbers have the same amount of utility as +/-. I don't want to sound like a dick but I'm just generally interested in if you have you read up on these stats at all?

 

[NobodyBeatsTheWiz]

Explain to me how Zone adjusted relative corsi numbers have the same amount of utility as +/-.
I don't want to sound like a dick but I'm just generally interested in if you have you read up on these stats at all?

Which zone adjusted relative corsi numbers?

 

[Burakovsky95*]

Which zone adjusted relative corsi numbers?

Obviously I'm not talking about the two games worth of data we have.

 

[g00n]

I definitely agree, but I hate seeing people getting flamed for using the stats we currently have. I don't think anyone is putting their 100% faith in what we have now and they shouldn't but dismissing them altogether like many do on this board is just stupid IMO.

Yeah, nobody ever puts 100% faith in them as "a completely proven analytical science". Or flames people who question the broad application of that faith.

http://hfboards.mandatory.com/showpost.php?p=90392165&postcount=93
http://hfboards.mandatory.com/showpost.php?p=90392351&postcount=98
http://hfboards.mandatory.com/showpost.php?p=90782465&postcount=103
http://hfboards.mandatory.com/showpost.php?p=90394855&postcount=122

Never happens. Especially not within the last week.