Ranking the top-40 goaltenders by Goals Allowed Per Expected Goal Against

[x Tame Impala]

The team you play on has a major effect in terms of how people view goaltending

Crawford was viewed as overrated since he was on stacked teams
Funny even though his team has turned for the worst, he's putting up respectable numbers. Now he's "underrated"...that's how the winds blow in fans' minds

I’ve had to read that bogus argument since 2013. Great players playing great makes a team great. The Hawks were so good for so long, and some years had a better record than they should’ve had, largely in part because Crawford was consistently very good to great.

All that says to me is that @JoVel doesn’t now and hasn’t prior watched many Hawks games.

 

[JaegerDice]

What is "expected GA"?? Sounds like one of those made up HFboards stats.

OP if you're going to use fake stats to make a ranking at least add your definition of it somewhere into your long post.

xGF and xGA (as well as xGF%) are all widely-adopted stats at this point. You can find how they're calculated with a quick google search.

 

[Sidney the Kidney]

xGF and xGA (as well as xGF%) are all widely-adopted stats at this point. You can find how they're calculated with a quick google search.

Asked this in a previous thread, but I was curious what the "point" of xGF (or the player equivalent, xG) stat is? Is it to predict what that player will do going forward assuming the same chances? And how accurate, historically, has it been in doing so?

For example, if a player scores 40 goals in a season but his xG stat is 25, does that mean it's suggesting he'll only score 25 goals the next season given the same sort of chances?

 

[JaegerDice]

Asked this in a previous thread, but I was curious what the "point" of xGF (or the player equivalent, xG) stat is? Is it to predict what that player will do going forward assuming the same chances? And how accurate, historically, has it been in doing so?

For example, if a player scores 40 goals in a season but his xG stat is 25, does that mean it's suggesting he'll only score 25 goals the next season given the same sort of chances?

Yeah, basically the model was built to try and improve on corsi as far as predicting future goals by introducing elements of 'shot quality'. Factors like distance, angle, shot type, deflections, if the shot was off the rush or off a cycle, if the shot was off a rebound, etc.

Expected Goals are a better predictor of future scoring than Corsi, Goals

Shot Quality And Expected Goals: Part I | Corsica

Shot Quality And Expected Goals: Part 1.5 | Corsica

 

[aufheben]

Asked this in a previous thread, but I was curious what the "point" of xGF (or the player equivalent, xG) stat is? Is it to predict what that player will do going forward assuming the same chances? And how accurate, historically, has it been in doing so?

For example, if a player scores 40 goals in a season but his xG stat is 25, does that mean it's suggesting he'll only score 25 goals the next season given the same sort of chances?

It’s one way to look at which players/teams are over-/underperforming, which you can view during the season or in retrospect. I’m sure it’s conception was to serve a more elaborate and specific purpose, but that’s generally why I check them out.

 

[TomasHertlsRooster]

Didn’t watch much of the post-season did you? Because that’s the only way I could see you came to the conclusion that he is a below average goalie.

In the playoffs, Rask would rank near the top by this metric.

Player	Team	GA per xGA (NST)	Ga per xGA (Corsica)*
Curtis McElhinney	CAR	0.66	0.65
Robin Lehner	NYI	0.77	0.77
Tuukka Rask	BOS	0.81	0.69
Ben Bishop	DAL	0.82	0.78
Philipp Grubauer	COL	0.91	0.86
Marc-Andre Fleury	VGK	0.97	0.82
Mike Smith	CGY	0.97	0.91
Matt Murray	PIT	0.98	0.83
Jordan Binnington	STL	1.01	0.95
Pekka Rinne	NSH	1.02	0.94
Sergei Bobrovsky	CBJ	1.06	0.99
Braden Holtby	WSH	1.07	0.99
Petr Mrazek	CAR	1.08	1.01
Martin Jones	S.J	1.08	0.97
Connor Hellebuyck	WPG	1.09	1.03
Frederik Andersen	TOR	1.10	1.05
Andrei Vasilevskiy	T.B	2.18	1.81

[TBODY] [/TBODY]

*Corsica's data set was missing all data from the Stanley Cup Finals. That is why I just ranked by the numbers from Natural Stat Trick.

No doubt, Rask had an excellent playoffs. However, it was pretty much out of the blue, and his regular season performances haven't been great. Think of his first 3 rounds as a hot streak and his Stanley Cup Finals performance as regression to the mean.

 

[Machinehead]

Didn’t watch much of the post-season did you? Because that’s the only way I could see you came to the conclusion that he is a below average goalie.

Going against four years of performance because you thought he was impressive in a couple of playoff games is exactly why people use analytics.

 

[nbwingsfan]

Going against four years of performance because you thought he was impressive in a couple of playoff games is exactly why people use analytics.

The same analytics that led you to believe McIlrath is better than Fowler and a future top 4D?

Yeah, I’ll keep using my eyes.

 

[Machinehead]

The same analytics that led you to believe McIlrath is better than Fowler and a future top 4D?

Yeah, I’ll keep using my eyes.

That wasn't analytics, that was me being stupid.

Now put that corpse back in the hole. It has bugs.

 

[Frank Drebin]

The same analytics that led you to believe McIlrath is better than Fowler and a future top 4D?

Yeah, I’ll keep using my eyes.

Lmao.

The same analytics that show that Mike Smith was better than Andrei Vaselivski or that Brian Elliot outplayed Carey Price over the past 3 seasons.

Even the biggest analytic supporters admit it's not an exact science so I don't know why it's so hard to admit that if the results are wack, the underlying statistic probably is too.

 

[Alexei Yashvalev]

I don't think it's outrageous to say that Halak (.922sv%) outplayed Rask (.912sv%) in the regular season this year. Rask was terrible to start the year.

 

[Machinehead]

Lmao.

The same analytics that show that Mike Smith was better than Andrei Vaselivski or that Brian Elliot outplayed Carey Price over the past 3 seasons.

Even the biggest analytic supporters admit it's not an exact science so I don't know why it's so hard to admit that if the results are wack, the underlying statistic probably is too.

My cat has outplayed Carey Price over the past 3 seasons.

He's 20th in sv% and 34th in GAA. That's "analytics" now?

 

[Machinehead]

I don't think it's outrageous to say that Halak (.922sv%) outplayed Rask (.912sv%) in the regular season this year. Rask was terrible to start the year.

These are some crazy analytics you're using.

 

[Frank Drebin]

My cat has outplayed Carey Price over the past 3 seasons.

He's 20th in sv% and 34th in GAA. That's "analytics" now?

Ok, so explain Vasilevski then. 4th in sv% and 6th in gaa.

Price is 11th out of 19 starters (min 145gp) for sv%. 8th/19 for GAA not great but still ahead of Brian Elliott.

Your cat (Lundqvist) was 15/19, and 18/19 for sv% and GAA.

 

[Machinehead]

Ok, so explain Vasilevski then. 4th in sv% and 6th in gaa.

Price is 11th out of 19 starters (min 145gp) for sv%. 8th/19 for GAA not great but still ahead of Brian Elliott.

Your cat (Lundqvist) was 15/19, and 18/19 for sv% and GAA.

Vasilevskiy played for the team that had arguably the best regular season in the history of the sport.

Lundqvist is a below league average goaltender. Has been since probably 15-16.

Also, you don't go from 34th to 8th in GAA by adding a minutes filter so I don't know where those numbers came from.

 

[Frank Drebin]

Vasilevskiy played for the team that had arguably the best regular season in the history of the sport.

Lundqvist is a below league average goaltender. Has been since probably 15-16.

Also, you don't go from 34th to 8th in GAA by adding a minutes filter so I don't know where those numbers came from.

NHL.com - Stats

Also, this is over the past 3 seasons, cumulative. Tbl missed the playoffs in 16-17

 

[TomasHertlsRooster]

Lmao.

The same analytics that show that Mike Smith was better than Andrei Vaselivski or that Brian Elliot outplayed Carey Price over the past 3 seasons.

Even the biggest analytic supporters admit it's not an exact science so I don't know why it's so hard to admit that if the results are wack, the underlying statistic probably is too.

The results aren’t wack, though. They’re better than most results you’ll get from ranking players by purely statistics.

Go to NHL.com and rank forwards by points per game over the past 3 seasons with an 82 game minimum. I guarantee you that list will be less accurate than this one.

 

[Frank Drebin]

The results aren’t wack, though. They’re better than most results you’ll get from ranking players by purely statistics.

Go to NHL.com and rank forwards by points per game over the past 3 seasons with an 82 game minimum. I guarantee you that list will be less accurate than this one.

These results, in my humble opinion, are wack.

 

[Machinehead]

NHL.com - Stats

Also, this is over the past 3 seasons, cumulative. Tbl missed the playoffs in 16-17

So Carey Price's middling numbers even when we filter out only guys who were starters all three seasons is the hill we're dying on?

Vasilevskiy also sucked that year. It's amazing how his numbers spiked with the best team in the league in front of him as opposed to a team that missed the playoffs.

 

[Machinehead]

HF: "ANALYTICS DON'T APPLY ANY CONTEXT!!!"

*Goaltending analytics are entirely based on context*

HF:

 

[Frank Drebin]

So Carey Price's middling numbers even when we filter out only guys who were starters all three seasons is the hill we're dying on?

Vasilevskiy also sucked that year. It's amazing how his numbers spiked with the best team in the league in front of him as opposed to a team that missed the playoffs.

So you're saying that you believe that Mike Smith, Cam Talbot and Jaro Halak have been better over the past three seasons than Rask, Vasi, and Price?

If your answer is anything other than "yes", what the hell are you doing defending this crazy stat/list?

If your answer is actually yes, please go on the record here and now stating so.

 

[TomasHertlsRooster]

NHL.com - Stats

Also, this is over the past 3 seasons, cumulative. Tbl missed the playoffs in 16-17

Why not compare apples to apples and use the same 5,000 minute filter?

NHL.com - Stats

Price ranks 18th with a SV% of .915. Mike Smith ranks 27th with a SV% of .911. They are not very far apart there. If the guy with a .915 was playing behind a very good defense, while the guy with a .911 was playing behind a very weak defense, I don't think it's crazy to suggest that the guy with a .911 was better. I mean, seriously, that is not a big gap.

Look at the shots price faced over the past 3 years:

[URL='https://ibb.co/w7kqRTW']

[/URL]

At even strength in particular, the high danger areas near the front of the net are all mostly blue, meaning he faces less than an average goalie from those areas. Meanwhile, the red spots, where he faces more than an average goalie, generally come from the perimeter. (The results on the PK are mixed.)

Now compare them to the shots that Mike Smith has faced:


That 2016-2017 Arizona defense in particular was meme worthy, and Smith posted a .914 SV% in front of them.

Can you not at least agree, that from what you see on these images, that the difference in the defense in front of them was clear as day? And can you not agree that if all else was equal, a goaltender posting the same exact numbers in Smith's skates would have performed better than a goaltender posting those numbers in Price's skates? Clearly, you can at least agree there, right? And you can agree that comparing the two side-by-side by save percentage, or even worse, goals against average, would be extremely unfair to Smith?

Now consider that their SV% is only .004% apart. For every 1,000 shots that they faced, Price stopped 4 more. That is 1 goal against per 250 shots, or roughly 1 extra goal against every 10 games.

Why is it so crazy to say that after adjusting for the defenses in front of them, Smith was actually a little better? You keep saying that it's crazy or "wack", but you haven't provided a reason as to why that is.

 

[Bleedred]

Going against four years of performance because you thought he was impressive in a couple of playoff games is exactly why people use analytics.

Remember where you are.

This is the land where the best way to judge a player is by what they've done lately, even if they were awful before their very recent surge.

Although, with Rask he seems to be rated more on what he did in his career up to 2015, but that playoff run this past Spring has also helped further fuel this notion that Rask is still an elite goalie the way he was through 2015.

 

[Machinehead]

So you're saying that you believe that Mike Smith, Cam Talbot and Jaro Halak have been better over the past three seasons than Rask, Vasi, and Price?

If your answer is anything other than "yes", what the hell are you doing defending this crazy stat/list?

If your answer is actually yes, please go on the record here and now stating so.

Vasilevskiy, I don't know. His numbers are outstanding but he seems to get hit harder than anyone by team impacts. He's a strange case.

Price and Rask have sucked the last three seasons, so yes. Rask, in particular, has been horrendous. Bad numbers on an exceptional defensive team. I'm honestly not sure what more Rask needs to do for people to stop thinking he's a great goaltender.

 

[Machinehead]

Remember where you are.

This is the land where the best way to judge a player is by what they've done lately, even if they were awful before their very recent surge.

Although, with Rask he seems to be rated more on what he did in his career up to 2015, but that playoff run this past Spring has also helped further fuel this notion that Rask is still an elite goalie the way he was through 2015.

It's so weird. Everyone openly acknowledges that it's the most volatile position ("goalies are voodoo!") but it's also the position where it takes like 7-10 years for opinions to change on a player, more so than any other position.