[GDT] Flames @ Abbotsford. 8pm MT. Calgaryflames.com

[Lanny_McDonald]

Originally Posted by Bingo

Your fear of simple event counting stats has always astounded me.

It's not a fear of them, it's actually understanding statistics, the importance of clean data, and how you model data to make inferences. As a result I find the methods of these advanced analytics to be inconsistent and unreliable, and any data grossly flawed. All statistics are based on simple event counting. The most reliable models feature consistent methods for classification and collection of data, where these advance statistics are a massive failure. All "events" are classified as equal, and I think we all know, all events are not equal. When you try and model bad data, all you get is a bad model. Take a stats class beyond the 100 level and you won't look at this stuff the same way again.

Originally Posted by PepsiFree

If advanced stats are WebMD then Lanny is an anti-vaxxer on YouTube

Actually, that would make me the exhausted healthcare professional who is fed up the COVIDiots, trying to explain to them that hydroxycloriquine and ivermectin are not good ways of battling the disease, and telling them to STFU and put on a mask. But I appreciate the effort you put in. Really earning that participation ribbon.

[Bingo]

Originally Posted by Lanny_McDonald

It's not a fear of them, it's actually understanding statistics, the importance of clean data, and how you model data to make inferences. As a result I find the methods of these advanced analytics to be inconsistent and unreliable, and any data grossly flawed. All statistics are based on simple event counting. The most reliable models feature consistent methods for classification and collection of data, where these advance statistics are a massive failure. All "events" are classified as equal, and I think we all know, all events are not equal. When you try and model bad data, all you get is a bad model. Take a stats class beyond the 100 level and you won't look at this stuff the same way again.

If that was true you'd be on to something.

It isn't.

Love how you finish it with "take a stats class" ... never change man, never change.

[Wastedyouth]

Originally Posted by Lanny_McDonald

It's not a fear of them, it's actually understanding statistics, the importance of clean data, and how you model data to make inferences. As a result I find the methods of these advanced analytics to be inconsistent and unreliable, and any data grossly flawed. All statistics are based on simple event counting. The most reliable models feature consistent methods for classification and collection of data, where these advance statistics are a massive failure. All "events" are classified as equal, and I think we all know, all events are not equal. When you try and model bad data, all you get is a bad model. Take a stats class beyond the 100 level and you won't look at this stuff the same way again.



Actually, that would make me the exhausted healthcare professional who is fed up the COVIDiots, trying to explain to them that hydroxycloriquine and ivermectin are not good ways of battling the disease, and telling them to STFU and put on a mask. But I appreciate the effort you put in. Really earning that participation ribbon.

Sweet, definitely come here and vent your frustrations on the community here. We appreciate it.

[Lanny_McDonald]

Originally Posted by Bingo

If that was true you'd be on to something.

It isn't.

But it is. It's been proven over and over. Data collection is inconsistent and does not follow a rigid methodology. Events are not equal. For example, shots on goal can be very different from each other. There are too many variables that make modeling the chaos of hockey virtually impossible.

Love how you finish it with "take a stats class" ... never change man, never change.

It wasn't a jab. It was a statement of truth. The more exposure you have to statistics, data collection, and data modeling, the quicker you come to understand the flaws in models. I just think that you would change your take on this subject if you had gone through more education in relation to stats and modeling. The simplistic view of event counting would be quickly dashed. It's why the adage of there are lies, damn lies, and statistics, is so relevant. Most people don't understand statistics or models, and don't realize it until they work through courses to develop that understanding. As a baseball guy and a stats junkie growing up, I thought I knew stats until I was forced to take five classes in stats to complete my terminal degree, then use that understanding to conduct research and prove consistency and validity. Now I look at stats very differently, especially when it comes to sports and politics. When methods are half baked and the result are models not consistent or able to predict anything.

Originally Posted by Wastedyouth

Sweet, definitely come here and vent your frustrations on the community here. We appreciate it.

That was a correction of the comparison that Pepsi made, nothing more. I was not calling anyone anything, so get a grip. If I believe someone is an idiot, I'll tell them that directly to their face.

[Bingo]

Originally Posted by Lanny_McDonald

But it is. It's been proven over and over. Data collection is inconsistent and does not follow a rigid methodology. Events are not equal. For example, shots on goal can be very different from each other. There are too many variables that make modeling the chaos of hockey virtually impossible.

No it's not.

CF is a shot attempt. They don't care from where.

SCF is a shot attempt from within a region that historically been where most goals have been scored from.

The clear definition of the two is pretty defendable if you move from one to the other.

HDCF is a shot attempt in the same region but coming on a rebound, deflection or pass.

Once again a clear definition of how you move from one to other giving you clarity.

Could they get better? For sure they could, but they're not the loosey goosey, undefined, undifferentiated mess you make them out to be.

They're just not. It's the exact rigid methodology you're calling for.

As I said they can get better, but in the meantime they present a pretty simple question in my mind. Would you rather have ...

a) more shot attempts than the other team?

For sure, but that may not get you the win as it doesn't define danger at all and really only partially describes who's carrying the play.

b) more scoring chances for?

For sure, more shots from in front and close to my goalie, makes me nervous, better to have more of those than the opposition.

c) more high danger scoring chances for?

For sure, rebounds, tips and passes into that zone are generally the ones that end up in the back of the net.

All of this is then formulated in xGF and is a model. You can argue that they weight things wrong or not how you'd weight them, but you can't get away from a/b/c above. Better to have more than less.

Originally Posted by Lanny_McDonald

It wasn't a jab. It was a statement of truth. The more exposure you have to statistics, data collection, and data modeling, the quicker you come to understand the flaws in models. I just think that you would change your take on this subject if you had gone through more education in relation to stats and modeling. The simplistic view of event counting would be quickly dashed. It's why the adage of there are lies, damn lies, and statistics, is so relevant. Most people don't understand statistics or models, and don't realize it until they work through courses to develop that understanding. As a baseball guy and a stats junkie growing up, I thought I knew stats until I was forced to take five classes in stats to complete my terminal degree, then use that understanding to conduct research and prove consistency and validity. Now I look at stats very differently, especially when it comes to sports and politics. When methods are half baked and the result are models not consistent or able to predict anything.

That's like saying no offence but you're wife is ugly.

It is a jab.

It's suggesting that only when I reach your heights of education and experience will I have the bright beacon of clarity and logic put upon me.

Sad you can't even see how arrogant you are.

[Textcritic]

Originally Posted by Lanny_McDonald

What do I have again? Common sense.

Ah, yes. "Common sense." The last bastion to hide behind for the Dunning-Kruger crowd.

I used to have a boss who had an old computer with a very noisy keyboard on his laptop, and my colleague and I used to make endless fun of him behind his back (and in front of his face) for the cacophony of ear-splitting typing he would produce at high speeds.

Whenever I read New Era's posts I picture this guy with his deafening keyboard, but then also screaming at his monitor while he hammers away.

[Rutuu]

Originally Posted by Textcritic

Ah, yes. "Common sense." The last bastion to hide behind for the Dunning-Kruger crowd.

I used to have a boss who had an old computer with a very noisy keyboard on his laptop, and my colleague and I used to make endless fun of him behind his back (and in front of his face) for the cacophony of ear-splitting typing he would produce at high speeds.

Whenever I read New Era's posts I picture this guy with his deafening keyboard, but then also screaming at his monitor while he hammers away.

Same guys that call for housing crashes.

[PepsiFree]

Originally Posted by Lanny_McDonald

Actually, that would make me the exhausted healthcare professional who is fed up the COVIDiots, trying to explain to them that hydroxycloriquine and ivermectin are not good ways of battling the disease, and telling them to STFU and put on a mask. But I appreciate the effort you put in. Really earning that participation ribbon.

No, my analogy was perfect. Your analogy is the perfect illustration of what Text said best:

Originally Posted by Textcritic

Ah, yes. "Common sense." The last bastion to hide behind for the Dunning-Kruger crowd.

Why do I know it’s perfect? Because you’re a layperson who took some stats classes 20 or 30 years ago talking about how you use “common sense” and how Bader is wrong because of “flaws” which you can’t even get right in your explanation of them.

Who is Byron Bader again? Oh right, he’s an actual data scientist who has his actual master’s of science with an actual statistics designation, has worked for actual companies in the field of analytics such as Shaw and FortisAB, has actually consulted on an NHL draft for an actual NHL team, and actually knows what he’s talking about.

You’re a nobody when it comes to analytics crying “I’m like a doctor!”…. a doctor who uses common sense and took some medical classes once, I guess.

Arrogance is cool, I kind of find it endearing when a guy or girl knows their #### tight and isn’t afraid to act like it. But you aren’t that, you haven’t earned that arrogance. Prove me wrong. I gave you Bader’s resume. Give me more than “uh I use common sense!!!” Why should anyone give you an once of respect you’ve yet to earn on the topic?

[DeluxeMoustache]

Originally Posted by Textcritic

Ah, yes. "Common sense." The last bastion to hide behind for the Dunning-Kruger crowd.

I used to have a boss who had an old computer with a very noisy keyboard on his laptop, and my colleague and I used to make endless fun of him behind his back (and in front of his face) for the cacophony of ear-splitting typing he would produce at high speeds.

Whenever I read New Era's posts I picture this guy with his deafening keyboard, but then also screaming at his monitor while he hammers away.

This does reveal a lot about you.

It appears you find loud typing somewhat funny. Cool

Funny enough that it is a reason to introduce a negative element to the work environment. (but not as funny as the comment about harem warfare, which along with the self description of acerbic cyber bully, seems to be part of your desired online identity)

Should I have the responsibility of a moderator, I imagine I would endeavour to picture people with more of a neutral lens. Also, in light of how people are increasingly noting the toxicity of the board, I would maybe shy away from posts that escalate conflict

I read his posts in the voice of Isaac Hayes. Tough to get worked up about that

That’s just me, though