2017-18 stats and underlying metrics thread [Mod: updated season]

[garret9]

Are you saying that you can predict a team's results 62% of the time, but 38% you are going to be lead astray by unevenly distributed variance ("luck")? Has this been tested?

I tend to get hung up on value and how it's derived. For example, saying shot quality drives 10% of the output. There's a lot of wiggle room on what a quality shot is, no?

I'm saying that the theoretical limit is 62%.
There are two methods to see this (one a thesis paper using machine learning and one a blog post comparison of win distribution) and they came to the exact same results (the only one that it was exactly the same in all significant digits). Both are linked in the article.

That's the theme for all of these estimates in the graphs; they are all verified from multiple people through multiple trials and experiments. There is some wiggle room, and the ratios will slightly change as the environment will change... but the basic relative relationship is what it is.

 

[Ippenator]

Honestly I don’t get it why some people have the need to try to predict how certain players will play in certain situations in the future. And then this predicting is even done with some completely irrelevant and meaningless arbitrary stats. What happened to the good old ”watch the games and comment after seeing how things really went”-approach?

In my opinion predictions are really a needless thing to do, unless you are going to make a money bet. And even that I wouldn’t do myself. I would rather put my money to funds or stocks that I have studied. Much better control for trying to win on a gamble in that way...

 

[YWGinYYZ]

Honestly I don’t get it why some people have the need to try to predict how certain players will play in certain situations in the future. And then this predicting is even done with some completely irrelevant and meaningless arbitrary stats. What happened to the good old ”watch the games and comment after seeing how things really went”-approach?

In my opinion predictions are really a needless thing to do, unless you are going to make a money bet. And even that I wouldn’t do myself. I would rather put my money to funds or stocks that I have studied. Much better control for trying to win on a gamble in that way...

Because some people enjoy the analysis part of hockey, in addition to watching it?

 

[Maukkis]

Honestly I don’t get it why some people have the need to try to predict how certain players will play in certain situations in the future. And then this predicting is even done with some completely irrelevant and meaningless arbitrary stats. What happened to the good old ”watch the games and comment after seeing how things really went”-approach?

In my opinion predictions are really a needless thing to do, unless you are going to make a money bet. And even that I wouldn’t do myself. I would rather put my money to funds or stocks that I have studied. Much better control for trying to win on a gamble in that way...

Sweet Jesus, what an abysmal take this is.

To answer the question in your post, I'd hazard the guess that stats are used to find out ways to improve the on-ice product. The rest of the post... oh my god. Stats are meaningless and irrelevant, but eye test works, right? Let me ask you this; which one of those means of evaluation let you "see what you want to see"?

 

[Ippenator]

Sweet Jesus, what an abysmal take this is.

To answer the question in your post, I'd hazard the guess that stats are used to find out ways to improve the on-ice product. The rest of the post... oh my god. Stats are meaningless and irrelevant, but eye test works, right? Let me ask you this; which one of those means of evaluation let you "see what you want to see"?

Both absolutely. Stats can be cherry picked and used for each posters own agenda as much as they care to twist them for their own purposes.

But only watching the games is really meaningful. Of course some real end results stats (like goals, assists, wins and losses) can for sure be interesting when thinking of how well the players have been doing. But they are no use for predicting anything for the future as any other stats aren’t as well.

Any kind of stat is just way, way too inaccurate when trying to predict things for individuals in different teams and constantly changing situations and attributes. Just a waste of time to speculate the future with stats that show only how things went in the past.

 

[grieves]

Both absolutely. Stats can be cherry picked and used for each posters own agenda as much as they care to twist them for their own purposes.

But only watching the games is really meaningful. Of course some real end results stats (like goals, assists, wins and losses) can for sure be interesting when thinking of how well the players have been doing. But they are no use for predicting anything for the future as any other stats aren’t as well.

Any kind of stat is just way, way too inaccurate when trying to predict things for individuals in different teams and constantly changing situations and attributes. Just a waste of time to speculate the future with stats that show only how things went in the past.

Everything in this world can be put in to stats. It's just a question of figuring out how to get enough information to form them. Stats are getting better and better. They can be extremely useful, and are only getting more useful as they are improved.

Some stats right now are pretty useless, but some definitely aren't.

 

[Ippenator]

Because some people enjoy the analysis part of hockey, in addition to watching it?

And what’s really the use of it? Fun to be wrong most of the time? I mean, even the most expertised hockey analysts are by a substantial margin much more wrong than right with their predictions. And before you start telling me something else, lets have a show of hands. Which ones of you can predict what happens in hockey accurately? And by accurately I mean that you can be that accurately and consistently right with your predictions, that you can invest money into different kinds of bets and get so consistently substantial wins that you can practically make a living with your predictions. I’m pretty sure this is a no show of hands after all...

 

[Ducky10]

And what’s really the use of it? Fun to be wrong most of the time? I mean, even the most expertised hockey analysts are by a substantial margin much more wrong than right with their predictions. And before you start telling me something else, lets have a show of hands. Which ones of you can predict what happens in hockey accurately? And by accurately I mean that you can be that accurately and consistently right with your predictions, that you can invest money into different kinds of bets and get so consistently substantial wins that you can practically make a living with your predictions. I’m pretty sure this is a no show of hands after all...

This is either deliberately obtuse or you don't really understand what statistics are used for and how. Statistics aren't used to make crystal balls, variance will always occur no matter what the situation. Stats give us probabilities, not guarantees. Nobody here is using stats to say what is going to happen, they are using them to say what is most likely to happen. In most cases, correctly.

 

[YWGinYYZ]

And what’s really the use of it? Fun to be wrong most of the time? I mean, even the most expertised hockey analysts are by a substantial margin much more wrong than right with their predictions. And before you start telling me something else, lets have a show of hands. Which ones of you can predict what happens in hockey accurately? And by accurately I mean that you can be that accurately and consistently right with your predictions, that you can invest money into different kinds of bets and get so consistently substantial wins that you can practically make a living with your predictions. I’m pretty sure this is a no show of hands after all...

It would be a boring world if we all had the same ideas and interests. Given that you've expressed your position on stats, I'm not at all interested in convincing you of their value in certain situations, what variances mean, etc.

 

[Maukkis]

Both absolutely. Stats can be cherry picked and used for each posters own agenda as much as they care to twist them for their own purposes.

But only watching the games is really meaningful. Of course some real end results stats (like goals, assists, wins and losses) can for sure be interesting when thinking of how well the players have been doing. But they are no use for predicting anything for the future as any other stats aren’t as well.

Any kind of stat is just way, way too inaccurate when trying to predict things for individuals in different teams and constantly changing situations and attributes. Just a waste of time to speculate the future with stats that show only how things went in the past.

If the bolded is true, then why do people have different opinions over plays that they have watched and seen happen with their own eyes? (hint: it isn't)

Stats may be deceptive, but there is one thing they have to their advantage: they are equal and never lie. What you are referring to as "inaccurate" is the people using those stats selectively to make a point. You know what? Those people are just as credible as those who rely on eye test to make theirs.

Also, going to leave this tweet here, since it may help you understand the ridiculousness of your post.

 

[Ippenator]

This is either deliberately obtuse or you don't really understand what statistics are used for and how. Statistics aren't used to make crystal balls, variance will always occur no matter what the situation. Stats give us probabilities, not guarantees. Nobody here is using stats to say what is going to happen, they are using them to say what is most likely to happen. In most cases, correctly.

And still they will be most of the times wrong if they claim how things will most likely go on the basis of stats. Life is so full of all kind of attributes that effect the end results, that it is honestly very naive to think that people can get hold of enough of the needed ones to get any useful predictions through historical stats. Sure you can calculate averages, but those have still nothing to do to with different individual players and all the different variables and attributes that have effects for each of them. Just waste of time all the more detail focused stats analytics. Sure nice, if some of you think that it is useful. But my stand is like this and nothing will change how I see this matter.

 

[Ippenator]

If the bolded is true, then why do people have different opinions over plays that they have watched and seen happen with their own eyes? (hint: it isn't)

Stats may be deceptive, but there is one thing they have to their advantage: they are equal and never lie. What you are referring to as "inaccurate" is the people using those stats selectively to make a point. You know what? Those people are just as credible as those who rely on eye test to make theirs.

Also, going to leave this tweet here, since it may help you understand the ridiculousness of your post.

Hint: stats always lie. They are no absolute truth. They are just the truth from some perspective, but have nothing to do with any universal absolute truth. There can always be found a stat that has a countering perspective and you always find stats that will support your agenda, if you have one. And none of them take into account all the different variables and attributes that change all the time for different individuals.

 

[Maukkis]

Hint: stats always lie. They are no absolute truth. They are just the truth from some perspective, but have nothing to do with any universal absolute truth. There can always be found a stat that has a countering perspective and you always find stats that will support your agenda, if you have one. And none of them take into account all the different variables and attributes that change all the time for different individuals.

This is your takeaway from my posts? Dear lord.

I guess I'll settle for utilising statistics that do not favour any individual player to form my opinions on players, as opposed to trusting the flawed human eyes that both you and I have. I can't really judge you for deciding to trust yours, though, but at least I'll know whose opinions I definitely won't hold in high regard.

 

[Gm0ney]

I've noticed there are plenty of fans who, like Han Solo, just don't want to know the odds.

 

[Ducky10]

And still they will be most of the times wrong if they claim how things will most likely go on the basis of stats. Life is so full of all kind of attributes that effect the end results, that it is honestly very naive to think that people can get hold of enough of the needed ones to get any useful predictions through historical stats. Sure you can calculate averages, but those have still nothing to do to with different individual players and all the different variables and attributes that have effects for each of them. Just waste of time all the more detail focused stats analytics. Sure nice, if some of you think that it is useful. But my stand is like this and nothing will change how I see this matter.

Learning is one of the greatest gifts in life, don't take it for granted.

Then again, ignorance is bliss, as they say.

 

[garret9]

Honestly I don’t get it why some people have the need to try to predict how certain players will play in certain situations in the future. And then this predicting is even done with some completely irrelevant and meaningless arbitrary stats. What happened to the good old ”watch the games and comment after seeing how things really went”-approach?

In my opinion predictions are really a needless thing to do, unless you are going to make a money bet. And even that I wouldn’t do myself. I would rather put my money to funds or stocks that I have studied. Much better control for trying to win on a gamble in that way...

If it's meaningless, irrelevant, and arbitrary, they won't predict anything and wouldn't last 10+ years and multiple studies and whatnot.

In my opinion, what ever happened to you know... having evidence to back up or double check an opinion? What's so bad with that...

 

[Ippenator]

Learning is one of the greatest gifts in life, don't take it for granted.

Then again, ignorance is bliss, as they say.

Yup, and believing in stats very much in predicting is exactly ignorance, ignorance of all the individually different attributes, variables and factors that can change so much that with each individual it is exactly ignorance to try to predict their future with all the very narrow and inaccurate information that any kind of stat watching gives.

Not really convinced that the great gift of learning is that often really used by a lot of people.

 

[garret9]

Both absolutely. Stats can be cherry picked and used for each posters own agenda as much as they care to twist them for their own purposes.

But only watching the games is really meaningful. Of course some real end results stats (like goals, assists, wins and losses) can for sure be interesting when thinking of how well the players have been doing. But they are no use for predicting anything for the future as any other stats aren’t as well.

Any kind of stat is just way, way too inaccurate when trying to predict things for individuals in different teams and constantly changing situations and attributes. Just a waste of time to speculate the future with stats that show only how things went in the past.

LOL this is hilarious.

Stats can be cherry picked, because anyone can bull shit. You can bull shit with what you see with watching the game too. There's whole fields of studies on how poorly the human memory and eye test performs in transcribing how events go.

Here's something though: stats that are scrutinized and peer reviewed do not say whatever you want. That's the difference.

If you found out that there was a way to tell how well a player was to outscore the opponent next year, that outperforms GMs opinions (in how they pay $) and how they did in outscoring (ie: goals)... how would you feel?

Stats is information. Any information can be misused if you intentionally ignore or over value information.
The eye test is also information.
Ain't nothing wrong with using as much information as possible.

 

[Ippenator]

LOL this is hilarious.

Stats can be cherry picked, because anyone can bull ****. You can bull **** with what you see with watching the game too. There's whole fields of studies on how poorly the human memory and eye test performs in transcribing how events go.

Here's something though: stats that are scrutinized and peer reviewed do not say whatever you want. That's the difference.

But the key is that no universal truth is needed for defining players. That’s the beauty of individuality. You can believe that stats prove really something about players. But in the end the only stats that mean anything are goals, wins and losses. All other stats are just arbitrary and take away the focus from the most important things in hockey.

In general people will anyway define the best players in their opinion with a set of different things. For some it is more about statistics, and for some it is more about the eye test. And for some it is a mixture of both. The variation with all this is individually different, and also the valuing of players is often a lot affected also by preferences with playstyles and as simple things as team based and nationality based bias even.

There is just absolutely no need or even possibility to find the absolute universal truth for which player is the best in any aspect in hockey. Or whom will play this well or that bad in the future on the basis of what he did in his past. And anyway, in the end the beauty is still going to be in the eye of the beholder. And that’s just how it will be, no matter if you like it or not.

 

[Aavco Cup]

Garret is right. There is a lot of BS around. Even right here in this very thread in fact. Although one would hope that scientific thread would discourage the BS somewhat....apparently not.

 

[garret9]

See, your post is a perfect example of how you can cherry pick information, using some truth, for an agenda even without using stats.
Now I'm going to use an example of how cherry picking doesn't make anything true because some will fail to scrutinization.

But the key is that no universal truth is needed for defining players. That’s the beauty of individuality. You can believe that stats prove really something about players.

There is a universal truth actually. This is objectively false because hockey has an quantitatively defined objective: score more goals than your opponent.

The universal truth is the best players help you reach this objective more effectively than others.

That's an irrefutable fact for as long as hockey stays being a sport where more goals wins. As long as the eye test holds value so will stats, because both of them will hold value for as long as some players can exhibit an impact on the game.

The only way statistics other than goals, assists, and wins cannot have value is if hockey is actually 100% random.

But in the end the only stats that mean anything are goals, wins and losses. All other stats are just arbitrary and take away the focus from the most important things in hockey.

Yes and no to the first sentence. No to the second sentence.

Wins and losses only matter as that's the objective: to garner the most wins.

Goals only matter because the definition of a win is the team with the most goals.

Other stats are not arbitrary as they, like the eye test, gather you information indicating which teams and players will garner you outscoring.

Some reading material that objectively shows you the non-arbitrary value of these statistics:
Objective NHL: Loose Ends - Part I: Predicting Future Success
Remembering Dellow: A few graphs to convince you on Corsi
Expected Goals are a better predictor of future scoring than Corsi, Goals

In general people will anyway define the best players in their opinion with a set of different things.

Two issues with this. People do not differently define what the best player represents, but what players represent the best player.

All people would agree that the best player gives you the best chance at winning. Figuring that out is filled with subjectivity, but that doesn't mean that some are more right than others.

For some it is more about statistics, and for some it is more about the eye test. And for some it is a mixture of both. The variation with all this is individually different, and also the valuing of players is often a lot affected also by preferences with playstyles and as simple things as team based and nationality based bias even.

Statistics and eye tests is not about defining what player is the best, but about trying to garner information to make an informed, estimated guess on which player is the best.

The destination, the path taken, and the efficiency of the path are not the same topics.

There is just absolutely no need or even possibility to find the absolute universal truth for which player is the best in any aspect in hockey. Or whom will play this well or that bad in the future on the basis of what he did in his past.

There absolutely is, and every single team and every single person in hockey will argue against this... otherwise their role has no real reason to exist.

And anyway, in the end the beauty is still going to be in the eye of the beholder. And that’s just how it will be, no matter if you like it or not.

Beauty is eye in the beholder in the subjective definition of which player you prefer. This is something about style over substance. Which player garners you the best chance at winning though is objective, and because certain players do indeed perform better than others there are patterns that can be detected.

 

[garret9]

Eye test is just a manner to gather information.
Statistics is just a manner to gather information.

Both methods have their strengths.
Eye test is superior for garnering information on why a player outputs the statistics they do.
Statistics is an objective manner to evaluate the impact a player has given and likely to give in the future.

Statistics and eye test will always matter provided:
* which team wins isn't random
* a team wishes to give them the best chance at winning

 

[winnipegger]

garret you have the patience of a saint. I'm not sure the target of your posts will fully appreciate the fullness of that response.

The fact that teams are paying number crunchers to help them tells you all you need to know about the utility of analytics. Money talks.

 

[Ippenator]

garret you have the patience of a saint. I'm not sure the target of your posts will fully appreciate the fullness of that response.

The fact that teams are paying number crunchers to help them tells you all you need to know about the utility of analytics. Money talks.

People buy tabloid magazines and spinner toys like crazy too. Doesn’t make either of those things really needed or important. Paid money is mostly nowadays more the proof of very well made marketing and sales than how good or needed the products really are. Sure part of it comes from the product often too, but still marketing and sales can be really amazing in selling whatever mumbojumbo nowadays.

Stats guys seems to be pretty good sales people and of course they have pretty much the ”what if”-factor that practically any NHL coach can’t neglect while they want to utilize even the smallest benefits they can get in the battle for the place in the sun

 

[Arthur Fonzarelli]

Yup, and believing in stats very much in predicting is exactly ignorance, ignorance of all the individually different attributes, variables and factors that can change so much that with each individual it is exactly ignorance to try to predict their future with all the very narrow and inaccurate information that any kind of stat watching gives.

Not really convinced that the great gift of learning is that often really used by a lot of people.

Do you realize you're arguing with someone that studies stats for a living and is paid for doing so by NHL teams?

On a side note I'd love to sit with you at a poker table.