How should we be judging goaltenders?

[Macbanan]

If you have to pick a single stat to judge a goalie the best is SV%. But as many have stated it needs context.
High number of shots inflate stats, low number deflate them.

Separating SV% by quality of scoring opportunity provides even more context.
It doesn't mean that HDSV% is any more important than LDSV%. It only means a low SV% can be explanied by a bigger ratio of high danger chances and vice versa.

Lastly all the context is available if you watch every game, but since we are very faulty beings I dont' trust our objectivity.
I trust my own take on goalies most of all and I know I'm not properly assessing all information, so I have to rely on stats as well.

 

[SladeWilson23]

Interesting stuff!

I wonder what it looks like if broken down into multiple tiers...possibly 35. Simple CA/60 search shows a range of 10 from first to worse over an entire decade so there certainly is significance there.

For this year......

25 OR FEWER SHOTS
GP: 208
SV%: .895
GAA: 2.39

26 TO 34 SHOTS
GP: 578
SV%: .916
GAA: 2.53

35 OR MORE SHOTS
GP: 354
SV%: .930
GAA: 2.70

 

[SladeWilson23]

If you have to pick a single stat to judge a goalie the best is SV%. But as many have stated it needs context.
High number of shots inflate stats, low number deflate them.

Separating SV% by quality of scoring opportunity provides even more context.
It doesn't mean that HDSV% is any more important than LDSV%. It only means a low SV% can be explanied by a bigger ratio of high danger chances and vice versa.

Lastly all the context is available if you watch every game, but since we are very faulty beings I dont' trust our objectivity.
I trust my own take on goalies most of all and I know I'm not properly assessing all information, so I have to rely on stats as well.

Perfectly said and couldn't agree more.

 

[pineapplestastegood]

The eye test. Save percentage is useless. Rask has the best save percentage in NHL history. Would anyone (rational) argue that he's the best goalie of all time? Of course not.

If there's some kind of save percentage on quality scoring opportunities, plus some kind of barometer for "weak ass goals allowed in clutch times to lose/tie the game", then maybe those could be statistics that would be worth something.

 

[I HATE THE PENS]

In the end, the only thing that matters is the team points, 2 for a win 1 for a ot loss, That is what it all comes down to. It is a team sport, The goalie can only do so much. So if you're team is in first place, You have a bad ass goalie.

 

[BrockH]

Deflate means a goalie can have a high SV%, but if they play in a lot of low shot volume games, sustaining it becomes more difficult. Expecting a goalie to sustain a high SV% over multiple seasons when playing on strong defensive teams is unrealistic. Marty Turco was able to do it for one season in 2003, but he never really came close again.

Since the 2005 lockout, when goalies face 29 or fewer shots, the average SV% is .905, and that's even after omitting the bad games where goalies get pulled early or come in for relief. When facing 30 or more shots, the average SV% is .925.

*Shrug* I feel like you're splitting hairs then by saying it's deflate not decrease because you have very clearly just stated that if a goalie has a high SV%, it is likely to decrease (since sustaining it becomes difficult) if they play in a lot of low shot volume games. Not arguing with your conclusions - and I really appreciate the tables you shared - but I don't think you need to worry so much about the semantics because deflate and decrease are virtually synonymous for your purposes, IMO.

 

[SladeWilson23]

*Shrug* I feel like you're splitting hairs then by saying it's deflate not decrease because you have very clearly just stated that if a goalie has a high SV%, it is likely to decrease (since sustaining it becomes difficult) if they play in a lot of low shot volume games. Not arguing with your conclusions - and I really appreciate the tables you shared - but I don't think you need to worry so much about the semantics because deflate and decrease are virtually synonymous for your purposes, IMO.

Perhaps you're right. I do however want to make sure there is a clear distinction.

I'm saying that one goalie who faces 30 shots a game is not guaranteed to have a higher SV% than a goalie who faces 26 shots a game.

 

[BrockH]

In the end, the only thing that matters is the team points, 2 for a win 1 for a ot loss, That is what it all comes down to. It is a team sport, The goalie can only do so much. So if you're team is in first place, You have a bad ass goalie.

Technically, the only thing that matters is Cups, so you're right...peak Osgood was a much better goalie than peak Hasek.

 

[SladeWilson23]

Technically, the only thing that matters is Cups, so you're right...peak Osgood was a much better goalie than peak Hasek.

I think what he's trying to say is that goalies have such a limited job description that you really only need a goalie who's "good enough" rather than need one who's great.

This by no means is me saying great goalies can't make a difference, because the goalie who was in my team's net for 19 seasons was absolutely the biggest reason for the 3 banners hanging.

 

[Drew Doubty]

No stat, or collection of stats, is going to give you a complete picture of how good a goalie is. The reason sv% is used is because it controls for shot volume and team quality better than GAA, wins, or shutouts. The counter to this is that good defensive teams suppress good shots and therefore the sv% of goalies on these teams will be inflated. Consistency is really what matters since any GM in the world would take a goalie that reliably delivers a .920 sv% over a guy whose percentage is all over the map, but consistency can only be established once someone has played 4 or 5 seasons in the NHL or even after their career is over. Stats will always be flawed.

The eye-test is always the best way to tell how good a goalie or any player really is. But we all can't watch all games so we must use statistics to aid our evaluation, and sv% is probably the best of the bunch.

 

[Stimpythecat]

This article seems like a good start for evaluations. Shot quality and expected goals against. It doesn't have data from this season so goalies like Darling, Price, and talbot are going to drop a ton while others like Quick and Vasilevskiy are going to rise.

Statistically Speaking: Goaltenders and expected goals against - Article - TSN

 

[toxic poster]

has anyone mentioned Cole Anderson (@CrowdScoutSprts) | Twitter here? Cole Anderson's work is the best in the world when it comes to goaltenders & his work specifically with rebounds & gsaa is phenomenal.

 

[SladeWilson23]

has anyone mentioned Cole Anderson (@CrowdScoutSprts) | Twitter here? Cole Anderson's work is the best in the world when it comes to goaltenders & his work specifically with rebounds & gsaa is phenomenal.

Excellent find!

 

[rent free]

wins

 

[powerbomb]

I tend to judge goalies by their helmet art. Sometimes talent alone isn't enough to separate between two skilled goaltenders, you know, but I can tell in a heartbeat who has the better mask.

 

[SladeWilson23]

Does anybody know how they calculate dSV% and xSV% on Corsica?

 

[Benneguin]

I tend to judge goalies by their helmet art. Sometimes talent alone isn't enough to separate between two skilled goaltenders, you know, but I can tell in a heartbeat who has the better mask.

That must mean 70s / 80s goalies >>>>>> Dave’s art era goalies.

 

[GoJetsGo55]

By how much their name rhymes with Smelly Truck.

 

[Seanaconda]

Eye test. All the other stats are skewed by what team they are playing for. But different styles also fit different systems .

Unless we get another hasek thats the best way I think

 

[sandysan]

SV% is not a good way to evaluate goalies who play on strong defensive teams.

Or weak defensive teams or teams who limit shots from high value areas, so pretty much the entire spectrum of all teams.

 

[sandysan]

so we should just look at stats that say somebody is good instead of watching a player and seeing that he’s good?

Im busy! I use spreadsheets all day. Who has time to watch games when there are regressions to the mean waiting to be done and relational databases to be mined to determine which goalie has the best save percentage on Wednesdays on a road trip after having given up 5 goals or more on the preeeding Monday.? People wanna know!

 

[Oddbob]

In a court of law, wearing a gown and with a gavel.

 

[Oddbob]

Im busy! I use spreadsheets all day. Who has time to watch games when there are regressions to the mean waiting to be done and relational databases to be mined to determine which goalie has the best save percentage on Wednesdays on a road trip after having given up 5 goals or more on the preeeding Monday.? People wanna know!

I wanna know who had the most 33 shot against nights, while playing with the flu!

 

[supsens]

This partially stems from a recent thread I saw in the Polls section about whether or not Robin Lehner was a #1 goaltender. The only argument I saw for Lehner from anyone in the entire thread was that his Sv% was good. I saw a couple Buffalo posters talk about how his Sv% numbers might be a bit misleading. Sure he absorbs a lot of shots, but does he make the key saves when he needs to? I'm really not trying to make this into a thread about Robin Lehner or any specific goaltender, this is purely an example.

right?

As a sens fan we had the same debate about Lehner cause his save % looked good,his suporters decided the team played better for Anderson then for him.
I personally thought the guy was a master of loosing by 1. Never getting blown out, but never making the save to hold the lead or the tie.
I know the thread isn’t about him but if the same debate is still going on, something must be funny with him.
Somewhat like Bobby lu being one of the best goalies in the world right up till the second you needed him to be, then he would just fall apart.
Some guys stop a ton of shots one night and suck the next. Some guys but very few are always hard to score on, and some guys let in one, two or three every game but manage to hang on and not give up the last goal for the loss.

 

[ulvvf]

save%.

wins is a pretty horrible way to judge goalies, like they would be the only player playing the game.

If you are that guy saying, oh well wins indicate importent saves? Well are not all saves importent? If you want to focus on the last minut of the game when everything is on the line kind of saves, well we can have save% stat of the last 2 minut of the games or something like then, instead of focusing on wins, which should not be on one players shoulders.

I know all saves is not the same, if you only have to take easy shoots then it makes your save% more impressing than it really is and vice versa. You could have a stat over advanced saves, but what is advanced saves, that becomes very arbitrarily.

No stats is perfect, but save% is by far the most telling stats when it comes to goalies, win or GAA do not come close. Then you can have different kind of save% to break it down, like last 2 minut save% etc, players on the rush save%, you can have all kinds of save%.