Proposal: Statistical Analysis

[91Kadri91*]

 

[91Kadri91*]

 

[91Kadri91*]

Here's a unique way of evaluating team 'momentum': http://www.stlouisgametime.com/2016/1/15/10777904/article

 

[Ovate]

Here's a unique way of evaluating team 'momentum': http://www.stlouisgametime.com/2016/1/15/10777904/article

Nobody would actually use that method to judge a stock and anybody who did would underperform the market, so I'm not sure why they'd think it's a useful tool for hockey analysis either.

Perfect example of the Maslow's Hammer mindset so common in hockey analytics.

 

[91Kadri91*]

Nobody would actually use that method to judge a stock and anybody who did would underperform the market, so I'm not sure why they'd think it's a useful tool for hockey analysis either.

Yeah, as I said, it's 'unique'.

Perfect example of the Maslow's Hammer mindset so common in analytics.

Fixed that for you.

 

[91Kadri91*]

Here's an article evaluating the two 'pick-for-picks' trades the Leafs made at the 2015 draft: https://mapleleafshotstove.com/2016/01/09/2015-draft-day-how-mark-hunter-and-kyle-dubas-may-have-outplayed-the-league/?utm_content=buffer7aa7c&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer

 

[Cap'n Flavour]

This is not how you fit a curve. He could have just used a moving average instead of fitting four points by eye.

 

[91Kadri91*]

Scoring Chances/Zone Data halfway summary: https://mapleleafshotstove.com/2016/01/14/toronto-maple-leafs-scoring-chances-and-zone-entries-summary-at-the-halfway-point/

 

[91Kadri91*]

Well, it appears that WarOnIce will not be around much longer: http://blog.war-on-ice.com/a-message-to-the-community/

The site will remain running until at least March 31, 2016, but we can’t guarantee that it’ll run after that. No new features will be added, but I expect those features that are under development by our other volunteers can easily live on elsewhere. There’s also no shortage of cap sites out there; we’ll continue to openly share all our cap info for anyone who wants to use it afterwards. But we will probably re-link our cap pages to one of the other sites at some point soon.

 

[91Kadri91*]

I have no idea what constitutes 'best', but:

 

[Menzinger]

Well, it appears that WarOnIce will not be around much longer: http://blog.war-on-ice.com/a-message-to-the-community/

Yep - not great news, but fortunately some other site will eventually pick up where they left off. It's in the nature of online communities.

I still miss Extra Skater though.

 

[91Kadri91*]

Credit to WilliamNylander for linking this article in the Rielly thread: http://www.sportsnet.ca/hockey/nhl/is-morgan-rielly-trending-to-be-a-no-1-defenceman/

 

[TheCLAM]

We should re-name this the Give 91Kadri91 Free posts thread

LOL "Statistical Analysis"

This thread is loaded with statistical errors passed off as a gold mine for innovation.

 

[Bomber0104]

We should re-name this the Give 91Kadri91 Free posts thread

LOL "Statistical Analysis"

This thread is loaded with statistical errors passed off as a gold mine for innovation.

No kidding.

It's beyond ridiculous.

 

[TheCLAM]

No kidding.

It's beyond ridiculous.

It might be time for the mods to step in lol

59 91Kadri91 posts / 139 total posts is about 42% of the entire discussion revolved around one poster.

Best part of waking up is Folgers in your cup

 

[Stand Witness]

We should re-name this the Give 91Kadri91 Free posts thread

LOL "Statistical Analysis"

This thread is loaded with statistical errors passed off as a gold mine for innovation.

No kidding.

It's beyond ridiculous.

It might be time for the mods to step in lol

59 91Kadri91 posts / 139 total posts is about 42% of the entire discussion revolved around one poster.

Best part of waking up is Folgers in your cup

What the hell are you two talking about? Want to argue the statistical errors? Why the hell do you care about how many posts he has in this thread? Ignore the thread if you don't like it. Not everyone is a cave man.

 

[Bomber0104]

What the hell are you two talking about? Want to argue the statistical errors? Why the hell do you care about how many posts he has in this thread? Ignore the thread if you don't like it. Not everyone is a cave man.

There aren't so many errors so much as it's simply bad data and a poor understanding of hockey and statistics at large.

There's a sub forum called By The Numbers and it's there for a reason.

 

[Stand Witness]

There aren't so many errors so much as it's simply bad data and a poor understanding of hockey and statistics at large.

There's a sub forum called By The Numbers and it's there for a reason.

And this is the Leafs sub forum with a thread dedicated to the underlying stats around the Leafs. I don't see why this is an issue at all.

If it is bad data and a 'poor' understanding of hockey and statistics then why don't you prove why it is like that and try to explain some reasoning behind it instead of saying "No kidding, it is beyond ridiculous". If you can't back up what you are saying with some tangible argument then there is no point in even saying it. If that is true you are quite honestly just trolling which no one wants to bother with.

 

[91Kadri91*]

Credit to WilliamNylander for linking this article in the Rielly thread: http://www.sportsnet.ca/hockey/nhl/is-morgan-rielly-trending-to-be-a-no-1-defenceman/

Here's Lindholm's chart:

 

[Ovate]

And this is the Leafs sub forum with a thread dedicated to the underlying stats around the Leafs. I don't see why this is an issue at all.

If it is bad data and a 'poor' understanding of hockey and statistics then why don't you prove why it is like that and try to explain some reasoning behind it instead of saying "No kidding, it is beyond ridiculous". If you can't back up what you are saying with some tangible argument then there is no point in even saying it. If that is true you are quite honestly just trolling which no one wants to bother with.

Honestly, a lot of these posts are blatantly wrong enough that anyone with a reasonable understanding of statistics would immediately dismiss them.

Unfortunately, imparting a reasonable understanding of statistics is difficult to do over a web forum, and impossible to do to someone who's argumentative and convinced that you're wrong.

Someone on the trade forums was saying that Laichs and Umberger are better value for their contract than Bozak. You can't explain to that kind of person why they're wrong, even though it's immediately obvious to almost anyone who looks at the comparisons.

 

[Stand Witness]

Honestly, a lot of these posts are blatantly wrong enough that anyone with a reasonable understanding of statistics would immediately dismiss them.

Unfortunately, imparting a reasonable understanding of statistics is difficult to do over a web forum, and impossible to do to someone who's argumentative and convinced that you're wrong.

Someone on the trade forums was saying that Laichs and Umberger are better value for their contract than Bozak. You can't explain to that kind of person why they're wrong, even though it's immediately obvious to almost anyone who looks at the comparisons.

People who like to use stats, numbers and facts to back up their arguments are the most open to being proven wrong. I enjoy debating when someone actually provides insight instead of a typical "you are wrong but I have nothing to back it up". If you can prove me wrong then do it and I will accept that but it has to be done with more than "you are wrong because I say so".

Prove something to me, don't just give me blank statements without anything supporting your argument.

 

[WilliamNylander]

Here's Lindholm's chart:

His and Rielly's are eerily similar. Wow.

 

[WilliamNylander]

We should re-name this the Give 91Kadri91 Free posts thread

LOL "Statistical Analysis"

This thread is loaded with statistical errors passed off as a gold mine for innovation.

Just stay out of the thread if you don't like it?

Some of us enjoy reading these posts, many of them are pretty cool/interesting.

 

[Ovate]

People who like to use stats, numbers and facts to back up their arguments are the most open to being proven wrong. I enjoy debating when someone actually provides insight instead of a typical "you are wrong but I have nothing to back it up". If you can prove me wrong then do it and I will accept that but it has to be done with more than "you are wrong because I say so".

Prove something to me, don't just give me blank statements without anything supporting your argument.

Consider the recent article that compared hockey analytics to the stock market, and then provided tools to poorly evaluate the stock market.

The reason that article was dead wrong wasn't that they did the math wrong. It was that they were making all kinds of wrong assumptions. You can't prove that with numbers. You just have to "get" stats and modelling and forecasting.

 

[Ovate]

Just stay out of the thread if you don't like it?

Some of us enjoy reading these posts, many of them are pretty cool/interesting.

People criticizing this thread aren't doing it to troll. The criticism is because we'd like to see actual intelligent discussion about stats instead of one member posting every single article he finds with little to no discussion or evaluation of the contents.