Normalizing shutouts

[Iain Fyffe]

Now if the Hawks were scouting the series what do they do? To read your analysis they say "Well 5 games are not sufficient or revealing, nothing to be learned here.".

That's a pretty massive shift of the goalposts. I thought we were discussing who was the better goaltender over their careers? Now you're talking about single-series micro-analysis, which is a tremendously different beast.

The Hawks rather obviously learned from the 5 game sample space and won the SC by going to the net at every opportunity.

Or perhaps they won the Cup because they were the #3 overall team playing against the #20 overall team? Leighton's luck ran out.

Besides, based on your prevous arguments I thought that Leighton's playoff shutouts indicate his true performance level against the best teams? Why would something as simple as going to the net be able to overcome his true ability in the playoffs?

 

[Canadiens1958]

NO...............

That's a pretty massive shift of the goalposts. I thought we were discussing who was the better goaltender over their careers? Now you're talking about single-series micro-analysis, which is a tremendously different beast.


Or perhaps they won the Cup because they were the #3 overall team playing against the #20 overall team? Leighton's luck ran out.

Besides, based on your prevous arguments I thought that Leighton's playoff shutouts indicate his true performance level against the best teams? Why would something as simple as going to the net be able to overcome his true ability in the playoffs?

No................... the issue was your characterization of shutouts as a garbage stat. I have shown repeatedly that it is far from a garbage stat but has many macro and micro applications in the historic analysis of NHL hockey.

Did I ever say that shutouts indicated Leighton's true performance? No.That you wish to put things in the context that you do is beyond my control

As for a massive shift of goalposts, well you have gone from attributing tremendous analytical powers to SV% and GAA or whatever is not shutout based to "Luck", which basically reduces a hockey game to the level of a lottery.

As for going to the net, the tactic forces the goalie to make multiple choices, reads, etc that he does not have to make when facing a perimeter game. The level of ability to make and execute these factors separates the NHL from AHL goalies. Going to the net focused on Leighton's biggest weakness which basically is the object of coaching and team preparation.

 

[Iain Fyffe]

No................... the issue was your characterization of shutouts as a garbage stat. I have shown repeatedly that it is far from a garbage stat but has many macro and micro applications in the historic analysis of NHL hockey.

You have failed to show that shutouts can show anything that other, better stats don't do already. For shutouts to have value by themselves, they must show something not captured by other stats.

Did I ever say that shutouts indicated Leighton's true performance? No.That you wish to put things in the context that you do is beyond my control

Here's the first thing you said about the value of the shutout stat:

"The meaningful part of the shutout stat is allowing zero goals.Knowing that the goalie does not flinch especially at key times."

How is this not a statement that shutouts indicate clutch performance?

As for a massive shift of goalposts, well you have gone from attributing tremendous analytical powers to SV% and GAA or whatever is not shutout based to "Luck", which basically reduces a hockey game to the level of a lottery.

That's a frequent misinterpretation when someone brings luck into the equation. Luck is a very important factor to consider when examing a short playoff series. Over longer periods, it's less important, because luck will even out over time. The vast majority of people fail to include luck into their analyses, and attribute it to things like wanting it more and the like. You cannot discount luck as a factor, but obviously it is not the only factor.

There is no stat that is completely independent of team performance. That doesn't mean we shouldn't avoid using the ones that are more reliant on team performance when we have better options available.

From least team-reliant to most team-reliant, we have:

1. Save percentage
2A. GAA
2B. Shutouts
4. W-L record

I've put GAA and shutouts as roughly equivalent because they both depend on both the number and the quality of shots faced. GAA is a better stat because it captures the goalie's performance in all of his games, not just in the games where he plays well.

As for going to the net, the tactic forces the goalie to make multiple choices, reads, etc that he does not have to make when facing a perimeter game. The level of ability to make and execute these factors separates the NHL from AHL goalies. Going to the net focused on Leighton's biggest weakness which basically is the object of coaching and team preparation.

And shutouts tell you absolutely nothing about that. If he had a 1.00 GAA, giving up 1 goal every game, against a perimeter game, would that somehow be less of an indicator than if he recorded a few shutouts?

Or let's say he had 3 shutouts but was bad in the other games, so his GAA in these games is 2.50. Do the shutouts then indicate anything about his overall performance, or are they just random variations within his performance?

If you're looking to see how he performs against a perimeter game, and you restrict your analysis to a small subset of said games you're going to mislead yourself. Why focus on the shutouts when you can look at his performance in all such games, by using save percentage for instance?

 

[Canadiens1958]

Again............

You have failed to show that shutouts can show anything that other, better stats don't do already. For shutouts to have value by themselves, they must show something not captured by other stats.


Here's the first thing you said about the value of the shutout stat:

"The meaningful part of the shutout stat is allowing zero goals.Knowing that the goalie does not flinch especially at key times."

How is this not a statement that shutouts indicate clutch performance?


That's a frequent misinterpretation when someone brings luck into the equation. Luck is a very important factor to consider when examing a short playoff series. Over longer periods, it's less important, because luck will even out over time. The vast majority of people fail to include luck into their analyses, and attribute it to things like wanting it more and the like. You cannot discount luck as a factor, but obviously it is not the only factor.

There is no stat that is completely independent of team performance. That doesn't mean we shouldn't avoid using the ones that are more reliant on team performance when we have better options available.

From least team-reliant to most team-reliant, we have:

1. Save percentage
2A. GAA
2B. Shutouts
4. W-L record

I've put GAA and shutouts as roughly equivalent because they both depend on both the number and the quality of shots faced. GAA is a better stat because it captures the goalie's performance in all of his games, not just in the games where he plays well.


And shutouts tell you absolutely nothing about that. If he had a 1.00 GAA, giving up 1 goal every game, against a perimeter game, would that somehow be less of an indicator than if he recorded a few shutouts?

Or let's say he had 3 shutouts but was bad in the other games, so his GAA in these games is 2.50. Do the shutouts then indicate anything about his overall performance, or are they just random variations within his performance?

If you're looking to see how he performs against a perimeter game, and you restrict your analysis to a small subset of said games you're going to mislead yourself. Why focus on the shutouts when you can look at his performance in all such games, by using save percentage for instance?

Again.................

Starting with luck. Your interpretation of luck evening out is akin to the mythical "Law of Averages". No such law, never existed, never will. You re invited to submit data that specifies when luck will even out. How much time? All that exists is the simple reality that success is a function of time - until someone figures out why a player is successful and changes the circumstances so that they are no longer successful. In the case of goalies shutouts are the ultimate success. Break it down from the ultimate to the circumstances when goals are scored and you will get somewhere. Do it from the average and you will rarely find the answer but you will get the luck excuse.

In terms of Leighton and the flinch factor my point is very consistent. Perimeter game he did not flinch, evidenced by the shutouts. Crash the crease game he flinched quite often, evidenced by the lack of shutouts.So the strategy answer is rather evident.


Again you are avoiding the key analytics,. Your points now center around luck, if, bad, etc. Hypotheticals and vagueness. Can you break it down into specifics. Quantify "bad". with a number be it GAA or SV%. In the alternative define a circumstance within the context of winning where a shutout would be "bad".

SV% is heavily team reliant under actual game conditions. Under shootout conditions it is not team reliant at all with certain credits to advance scouting or game film study.. Under actual game conditions, SV% is a function of advanced scouting and its application into a team strategy that helps the goalie perform to his optimum.

 

[Iain Fyffe]

Starting with luck. Your interpretation of luck evening out is akin to the mythical "Law of Averages". No such law, never existed, never will.

What people think of as the Law of Averages is actually called the Law of Large Numbers. But the Numbers in that Law are typically too large to be of any direct use in hockey analysis.

So there is not a strict application of such a law. But if you believe that luck is no less important over 82 games than it is over 7, I'm not sure what to say.

In terms of Leighton and the flinch factor my point is very consistent. Perimeter game he did not flinch, evidenced by the shutouts. Crash the crease game he flinched quite often, evidenced by the lack of shutouts.So the strategy answer is rather evident.

Replace all instances of "shutout" in the above with "save percentage". If he has a lot of shutouts, he'll also have a high save percentage. How do the shutouts tell us anything that save percentage does not already tell us? You still haven't answered that question.

Leighton had 3 SO against Montreal in 2010. But he also had a save percentage of .950. What information do the shutouts provide that the extremely high save percentage does not?

He could have had 3 SO because the Philly defence completely contained the Habs forwards, and only allowed 10 shots per game. He could have had a crappy save percentage and still recorded the shutouts, if his team was dominating defensively. By looking at the save percentage, you avoid that trap.

Again you are avoiding the key analytics,. Your points now center around luck, if, bad, etc. Hypotheticals and vagueness. Can you break it down into specifics. Quantify "bad". with a number be it GAA or SV%. In the alternative define a circumstance within the context of winning where a shutout would be "bad".

I have no idea what you're asking me here.

Under actual game conditions, SV% is a function of advanced scouting and its application into a team strategy that helps the goalie perform to his optimum.

And this applies to shutouts threefold (as an estimate). Anything that serves to increase save percentage also serves to increase shutouts, with the additional factor that shutouts are affected by not only the quality of the shots faced but the number of shots faced as well.

Now, do you want to get back to Plante and Hall, which is where we started off? Or do you concede the point that Plante's higher shutout rate in the playoffs is not proof that he was a much better goaltender than their relative regular-season performances would suggest?

 

[seventieslord]

Iain, don't play his game...

 

[Iain Fyffe]

Iain, don't play his game...

Good advice, I think.

 

[Canadiens1958]

Luck etc

What people think of as the Law of Averages is actually called the Law of Large Numbers. But the Numbers in that Law are typically too large to be of any direct use in hockey analysis.

So there is not a strict application of such a law. But if you believe that luck is no less important over 82 games than it is over 7, I'm not sure what to say.


Replace all instances of "shutout" in the above with "save percentage". If he has a lot of shutouts, he'll also have a high save percentage. How do the shutouts tell us anything that save percentage does not already tell us? You still haven't answered that question.

Leighton had 3 SO against Montreal in 2010. But he also had a save percentage of .950. What information do the shutouts provide that the extremely high save percentage does not?

He could have had 3 SO because the Philly defence completely contained the Habs forwards, and only allowed 10 shots per game. He could have had a crappy save percentage and still recorded the shutouts, if his team was dominating defensively. By looking at the save percentage, you avoid that trap.


I have no idea what you're asking me here.


And this applies to shutouts threefold (as an estimate). Anything that serves to increase save percentage also serves to increase shutouts, with the additional factor that shutouts are affected by not only the quality of the shots faced but the number of shots faced as well.

Now, do you want to get back to Plante and Hall, which is where we started off? Or do you concede the point that Plante's higher shutout rate in the playoffs is not proof that he was a much better goaltender than their relative regular-season performances would suggest?

Luck as you are trying to use it is another word for a failure , unwillingness or not knowing how to prepare.

Step-up and quantify "bad" be it with SV% over/under, GAA over/nder. You cannot do it. But you cannot define a shutout in terms of "bad".

Point of fact is that in the first two Leighton shutouts the Canadiens outshot the Flyers, difference being the provenance and quality of the shots.

Plante was vastly superior to Hall and the playoff data hilited by shutouts clinch the issue. Even if you look at their St.Louis playoff numbers with the same team in front of them.

 

[Bear of Bad News]

Luck as you are trying to use it is another word for a failure , unwillingness or not knowing how to prepare.

Step-up and quantify "bad" be it with SV% over/under, GAA over/nder. You cannot do it. But you cannot define a shutout in terms of "bad".

Point of fact is that in the first two Leighton shutouts the Canadiens outshot the Flyers, difference being the provenance and quality of the shots.

Plante was vastly superior to Hall and the playoff data hilited by shutouts clinch the issue. Even if you look at their St.Louis playoff numbers with the same team in front of them.

You're still dodging the questions. They aren't hard questions.