over a period of time, all shots are equal in quality? This is absolutely not true. This is a weak way of avoiding the fact that the quality of shots are damn near immeasurable and therefore are ignored. Have you watched an Oilers game this year?
- Jeropotato

You said it.

I know you like the "scientific" shots are all the same theory but wouldn't a stat that rewards points based on shot quality be an extremely useful tool when evaluating a player? .5 points for a wrist shot from the slot, .25 for a wrist shot from the blueline or something similar to that.
- Dangles13

It probably would be and people are doing interesting things with shot quality so that they can use these stats to be more effective in the short term. But over the long term, it wouldn't make a difference.

Thanks, it makes some sense but has anyone ever tested the theory? I would be really interested to see time of possession stats and see the relation between it and Corsi. Hockey fans know that you can be chasing the puck all game and generate quick shots off turnovers which is a good thing for Corsi but doesnt actually reflect you didnt have the puck much. I see it with 4th lines forechecking well and generating a few good shots but dont have the hands to handle the puck for very long
- WhiteLie

Going back to 2005 when the NHL first starting collecting shot-block data, Corsi has proven to be the most effective predictor of future events of all stats in hockey and it has been tested tons of times by tons of different people.

It probably would be and people are doing interesting things with shot quality so that they can use these stats to be more effective in the short term. But over the long term, it wouldn't make a difference.
- James_Tanner

yes it would .

No, the fundamental criticisms are not the same at all. You would need need ten games of goal data to get one game of shot data. Also, over enough time, all shots are equal in quality, where as, until you have thousands upon thousands of goals you can't say the same thing. So, if you track goals - like plus minus does - you get skewed and unusable data that is mostly garbage.

If you think this sounds crazy, keep in mind that virtually all NHL teams now have analytics departments and the correlations and accuracy of the claims have been vetted by math professionals.

I can "run my mouth" about this all day because its true, accurate and factual.

Plus, by Google it, I meant that you should pull up a search engine and type your queery in there because then you could read one of the thousands upon thousands of posts fully and thoroughly discrediting plus minus and save me the hassle.
- James_Tanner

Bahahaha man you really don't get it do you??? What do you think I meant when I said the error gets scaled? Other than that scaling factor (which you don't even understand and it's not a simple 10:1 or whatever that garbage you just said was) pretty much all of the criticisms are of the same fundamental nature.

Hence why you come across like a clown when you say one is so important and the other is meaningless. Both can be very misleading because just as Jero said there is so much more going on that isn't even being accounted for.

It probably would be and people are doing interesting things with shot quality so that they can use these stats to be more effective in the short term. But over the long term, it wouldn't make a difference.
- James_Tanner

How long term?

yes it would .
- Jeropotato

Hey, if you can prove this to me, I would be open because as someone who considers themselves to be an intelligent person, there is nothing in this world that I won't change my mind about when shown new information.

But I have had this concept patiently explained to me, and proven with examples,charts etc by a person who I trust is qualified to do so because he is a real life math teacher and hockey writer (Steve Burtch).

He could be disputed and he could be wrong, but I am yet to see proof.

I don't know people have such a problem with averaging data, it's a very useful and mathematically sound concept.
- James_Tanner

Not the way you are using it. I think you need to read The Black Swan by Nassim Taleb because you clearly need a reality check on how you apply stats.

How long term?
- Dangles13

Hard to say, I am not an expert. I just go by what I know until I learn more. But I would estimate that after ten games you have 300 shots to go by, so that's proabably a decent ballpark guess.

David Clarkson.
It should be who gets paid the most per point.

I know you're being sarcastic, but the data is irrefutable.
- James_Tanner

Man, you are something else

Hard to say, I am not an expert. I just go by what I know until I learn more. But I would estimate that after ten games you have 300 shots to go by, so that's proabably a decent ballpark guess.
- James_Tanner

So over ten games shot quality doesn't matter?

Not the way you are using it. I think you need to read The Black Swan by Nassim Taleb because you clearly need a reality check on how you apply stats.
- RonPielep

Feel free to elaborate and explain, if it sounds good I'll check it out. I didn't get to be the smartest owl in the forest my ignoring new information.

David Clarkson.
It should be who gets paid the most per point.
- FLflames34

So over ten games shot quality doesn't matter?
- Dangles13

I don't think it ever really matters to be honest. Maybe in one game where its a total anomaly.

Man, you are something else
- Lee Denault

Sorry there Gramps, but the data that Corsi corresponds to winning is irrefutable. Its the best predictor of who will and won't win that we have. Something might come along later, but right now, it is.

Hey, if you can prove this to me, I would be open because as someone who considers themselves to be an intelligent person, there is nothing in this world that I won't change my mind about when shown new information.

But I have had this concept patiently explained to me, and proven with examples,charts etc by a person who I trust is qualified to do so because he is a real life math teacher and hockey writer (Steve Burtch).

He could be disputed and he could be wrong, but I am yet to see proof.
- James_Tanner

Is someone counting how many shots of certain velocity?
Is someone counting how many shots were uncreened?
Is someone counting how far from the shooter was from the goalie?
Screened shots? Tip ins? Good Goalie? Bad Goalie?

There's a million variables that actually dont even out. Dallas Eakins ran a quality corsi team that lost every game. Coach Nelsons team has taken a dip in Corsi but they are winning more?

Do you know why? I'll give you a hint, it isnt luck.

I don't think it ever really matters to be honest. Maybe in one game where its a total anomaly.
- James_Tanner

Then Corsi is pretty useless to a coach, what's he going to do? Walk into the room and say "Hey guys, if we don't let them shoot then they probably won't score."?

Bahahaha man you really don't get it do you??? What do you think I meant when I said the error gets scaled? Other than that scaling factor (which you don't even understand and it's not a simple 10:1 or whatever that garbage you just said was) pretty much all of the criticisms are of the same fundamental nature.

Hence why you come across like a clown when you say one is so important and the other is meaningless. Both can be very misleading because just as Jero said there is so much more going on that isn't even being accounted for.
- RonPielep

This is just ridiculous. Talk about a sophistic argument. The entire advanced stats community disregards plus minus - in fact, in 2015 virtually all NHL people do. Anyone who uses it gets instantly and mercifully mocked (not saying thats right or ok, but it is true).

They also uniformly agree that Corsi is the most telling stat when it comes to things like predicting who might make the playoffs from here on out. (Bad possession teams almost always lose over time when their hot goaltending or high shooting percentages run out).

So, sure, I am a clown. But I guess that means that every single analytics based writer and analyst is also a clown. Good to know.

Feel free to elaborate and explain, if it sounds good I'll check it out. I didn't get to be the smartest owl in the forest my ignoring new information.
- James_Tanner

Basically when statisticians don't understand risk and error and they decide to use Gaussian distributions and Central Limit Theorem to predict that these risks are negligible and confirm to society that all is safe but in reality they are pretty (frank)ing clueless and then 2008 happens and everyone goes...WTF?

Sorry there Gramps, but the data that Corsi corresponds to winning is irrefutable. Its the best predictor of who will and won't win that we have. Something might come along later, but right now, it is.
- James_Tanner

did you know that the leafs were playing better corst after carlyle got fired?

lost 23 of the last 27 games

did you know the leafs were playing terrible corsi before carlyle was fired?

6th place in east.

Is someone counting how many shots of certain velocity?
Is someone counting how many shots were uncreened?
Is someone counting how far from the shooter was from the goalie?
Screened shots? Tip ins? Good Goalie? Bad Goalie?

There's a million variables that actually dont even out. Dallas Eakins ran a quality corsi team that lost every game. Coach Nelsons team has taken a dip in Corsi but they are winning more?

Do you know why? I'll give you a hint, it isnt luck.
- Jeropotato

Sorry, I like and respect you , but you're fundamentally wrong here.

This is just ridiculous. Talk about a sophistic argument. The entire advanced stats community disregards plus minus - in fact, in 2015 virtually all NHL people do. Anyone who uses it gets instantly and mercifully mocked (not saying thats right or ok, but it is true).

They also uniformly agree that Corsi is the most telling stat when it comes to things like predicting who might make the playoffs from here on out. (Bad possession teams almost always lose over time when their hot goaltending or high shooting percentages run out).

So, sure, I am a clown. But I guess that means that every single analytics based writer and analyst is also a clown. Good to know.
- James_Tanner

nothing ironic about stats guys thumbing their noses at other stat categories.

This is just ridiculous. Talk about a sophistic argument. The entire advanced stats community disregards plus minus - in fact, in 2015 virtually all NHL people do. Anyone who uses it gets instantly and mercifully mocked (not saying thats right or ok, but it is true).

They also uniformly agree that Corsi is the most telling stat when it comes to things like predicting who might make the playoffs from here on out. (Bad possession teams almost always lose over time when their hot goaltending or high shooting percentages run out).

So, sure, I am a clown. But I guess that means that every single analytics based writer and analyst is also a clown. Good to know.
- James_Tanner

the analytics guys that beat the drum as Corsi being the tell all stat are indeed dumb. the ones that use it effectively as a tool or aid are pretty smart though.

Sorry, I like and respect you , but you're fundamentally wrong here.
- James_Tanner

how is he wrong?

All he is saying is there are many different variables and lists them.

How is that fundamentally wrong?