Since 1989-90... give me your best, peak condition, Team Canada, Team USA, or Team Europe

[Michael Farkas]

I think you think you're saying something here, but of course not, skaters (especially 5 on 5) play a free-flowing game, up and down, sometimes a team has possession a lot longer, some teams create more dangerous opportunities. So we don't care as much about "averages" because doing things like controlling possession, creating more chances, very valuable. How does that relate to goaltenders?

Don't think you think you know what I think to think for me...

This isn't my point. This is yours. "What is a goalie but averages"

The answer is: everything. I'm just trying to figure out how you came to your conclusion.

yes, this is how small sample sizes work. Sam Gagner had 8 points one game, that doesn't make him the best player ever at peak value because of that one game

Context matters. The query was about playoff series (which, I find important). A single game extreme example doesn't help this discussion. Or put another way, "I think you think you're saying something here, but of course not..." haha

Nobody said it was...

Well, perhaps there isn't documentation to that effect...but reading the room, it's quite clear that it's held in the highest regard. Unless it's not...in which case, how did we get into this discussion in the first place...?

So one rate stat for a different one..?

Ok, so not save % or GAA, but the "Michael Farkas Certified Bad Goal Against That Really Lets the Air out of a Balloon"... that's still an averaging/rate stat. If you don't like save %, fine, I doubt anyone considers a certified end all be all, sort by SV%, call it a day and be done. Whether you use a real or hypothetical different measure, you're still measuring a goaltender by a rate, based on shots faced, minutes played, you have to. It's the nature of the position in the sport they play. In baseball, we measure everything by rate, because how else can you? By plate appearances or pitches thrown just shows how much a manager believed in you/how they viewed you compared to alternative options, but it says nothing about actual performance. Similar to a goaltender. You can make a note if a guy started 65 games a bunch of years in a row and the implication of that, but the performance itself is going to be rate-based.

Yeah, I mean, you can frame it however you like. The notion that it would be bad to contextualize numbers with evaluation is...interesting. If it becomes another "rate stat", is that bad? You haven't seen it publicly...what if this rate stat correlates to winning way more often than the other two existing stats? What if it correlated with whatever your talent evaluation process for goalies is (if any)?

I'm not vehemently against...math...or types of math...I don't know how else to address your concerns to be honest...

It'd be relatively easy? Interesting theory... So you only look at the goals given up and not the save made? Why not both?

I do look at both. Who says that I don't? But what's more likely for someone like you to do: evaluate every developing offensive opportunity against OR evaluating the two or three goals that a goalie gives up a night?

I'm not trying to put the evaluation process on some nebulous, unreachable pedestal...quite the opposite. I want hockey fans to be better at seeing the game well. And part of that is, where possible, make it as approachable as you can.

So if I said..."you can't just look at saves, you have to look at every developing offensive zone opportunity against" well, who has time for that? So a reasonable person would brush me off and tell me to get lost...and they're right.

But if we make it bite sized - and I've personally done this, some private analytics groups do this - by just evaluating what's easily available, you get a little closer to where you want to be.

For one, it's a measure of how effectively he was in fact stopping pucks. If you think a 5 save shutout is the same as a 50 save shutout, that's a different one for me.

There isn't enough information there to render a decision, as extreme of an example as this is.

 

[WarriorofTime]

Don't think you think you know what I think to think for me...

This isn't my point. This is yours. "What is a goalie but averages"

The answer is: everything. I'm just trying to figure out how you came to your conclusion.

Be more specific "everything", what is everything and how is it not in some way, an average?

Context matters. The query was about playoff series (which, I find important). A single game extreme example doesn't help this discussion. Or put another way, "I think you think you're saying something here, but of course not..." haha

Ok, 4 games instead of 1 game, what point fundamentally changes..?

Well, perhaps there isn't documentation to that effect...but reading the room, it's quite clear that it's held in the highest regard. Unless it's not...in which case, how did we get into this discussion in the first place...?

Ok, so nobody disagrees with anyone...

Yeah, I mean, you can frame it however you like. The notion that it would be bad to contextualize numbers with evaluation is...interesting.

Interesting strawman...

If it becomes another "rate stat", is that bad? You haven't seen it publicly...what if this rate stat correlates to winning way more often than the other two existing stats? What if it correlated with whatever your talent evaluation process for goalies is (if any)?

Your hatred was of rate stats because we don't evaluate forwards that way...

I'm not vehemently against...math...or types of math...I don't know how else to address your concerns to be honest...

I do look at both. Who says that I don't? But what's more likely for someone like you to do: evaluate every developing offensive opportunity against OR evaluating the two or three goals that a goalie gives up a night?

I can't say your method of "who gives up the least bad goals" is a better method, I generally look at who gives up the least bad AND prevents the best, an intersection if you will... but maybe there's data to support the air out of balloon theory?

I'm not trying to put the evaluation process on some nebulous, unreachable pedestal...quite the opposite. I want hockey fans to be better at seeing the game well. And part of that is, where possible, make it as approachable as you can.

No offense, but now you're just being a bit patronizing. Is seeing the game well = seeing how you see it?

So if I said..."you can't just look at saves, you have to look at every developing offensive zone opportunity against" well, who has time for that? So a reasonable person would brush me off and tell me to get lost...and they're right.

But if we make it bite sized - and I've personally done this, some private analytics groups do this - by just evaluating what's easily available, you get a little closer to where you want to be.

Perhaps your secret private analytics group will allow the public to see their methods one day then.

There isn't enough information there to render a decision, as extreme of an example as this is.

Sure maybe all 5 saves were amazing (as improbable as that may be) and all 50 saves were the equivalent of stopping a beach ball (as improbable as that may be). I guess we can say that about every aspect of anything then, but we do sometimes use data/statistical probability to make determinations that we don't have time for our eyesight to catch, or even if we did, to freely remove our cognitive biases.

 

[Dingo]

save percentage leaders each season, minimum 41 games played:

Hellebuyck
Ullmark
Shesterkin
Varlamov (shortened season, 36 games played)
Rask
Bishop
Raanta (eek)
Bobrovsky
Elliot
Price
Rask !!!!!!!!
Bobrovsky !!!!!!! (shortened season)
Smith/Lundqvist
Thomas
Rask THREE
Thomas
Ellis (eek)
Backstrom
Hasek
Roloson
Turco
Theodore
Dunham (eek?)
Belfour
Hasek !!!
Hasek THREE
Hasek FOUR
Hasek FIVE
Hasek SIX (shortened season, still played 41)
Hasek SEVEN
Joseph
Roy
Belfour !!!!
Roy !!!
Roy THREE
Roy FOUR
Hextall
Froese
Lindbherg
Smith
Melanson
Fuhr/Smith (Smith TWO?)
Edwards
Esposito
Resch
Dryden
Dryden TWO
Resch TWO
Vachon
Parent
Dryden THREE
Esposito TWO

stopping here. Tretiak should have been in that last mix of goalies. Perils of hockey pre-1989.

So, to your point - Boston had two goalies at the same time, as did Philly in the mid 80s (three actually) and NYI early 80s. Point conceded, team defence has shown itself. I will say that Rask also led much later on with what wasnt the SAME team. I do think he has been an elite goalie, and Thomas DID take a young Rask's spot. But, point conceded.

Now, there are some blips on this list, can you concede that it is largely very good goalies?

and, to my theory of the luck of lightning striking twice for a less than stellar goalie -

Hasek 7
Roy 4
Dryden 3
Rask 3
Belfour 2
Esposito 2
Resch 2
Thomas 2
Bobrovsky 2 (kinda, shortened season)
Billy Smith 2 (kinda, one shared)

Like, thats a pretty skookum list.

edit - if snyone wants to know where Brodeur is - probably very high on the list of GAAchampions, as that IS a team stat, influenced by goalies, whereas save% is a goalie stat, influenced by team.

 

[Kaiden Ghoul]

Mcdavid Gretzky Lemieux
Mackinnon Lindros Iginla
Yzerman Crosby Bergeron
Messier Sakic St.Louis

Pronger Bourque
McInnis Makar
Coffey Stevens

Roy
Brodeur

 

[Dingo]

Mcdavid Gretzky Lemieux
Mackinnon Lindros Iginla
Yzerman Crosby Bergeron
Messier Sakic St.Louis

Pronger Bourque
McInnis Makar
Coffey Stevens

Roy
Brodeur

thsnks for retracking us. thats a good team.

 

[WarriorofTime]

save percentage leaders each season, minimum 41 games played:

[list of really great goalies]

That's nice and all, but we still need to see the "Michael Farkas Certified Bad Goal Against That Really Lets the Air out of a Balloon" to really understand how bad Tim Thomas actually was and how SV% is a conspiracy

 

[Michael Farkas]

Be more specific "everything", what is everything and how is it not in some way, an average?

With your question. you highlight a point that I've seen from you and there's a fundamental difference in how we view the game. I don't view the game from the perspective of "stats...and in". I don't watch a goalie and go, "hmm...I wonder what his average is?"

I'm looking at his process. What is his process for making a save at the level he's at?

Ok, 4 games instead of 1 game, what point fundamentally changes..?

4 games are required to win a playoff series today. One regular season game is N/A.

Ok, so nobody disagrees with anyone...

Interesting strawman...

Wait, do we have to do the internet kangaroo court stuff...? I already saw it in the last post and discarded it "strawman" and "appeal to authority" and all these buzzwords...does it have to go that way? I don't care about that kind of stuff. You win all those points, whatever they are...

In any case, again, that wasn't my point...it was yours. I don't know how else to express it than that. You framed me up for something and I said, "no, that's not how it would work...why would it?" haha

Your hatred was of rate stats because we don't evaluate forwards that way...

That's not a fair representation of my argument or belief.

I can't say your method of "who gives up the least bad goals" is a better method, I generally look at who gives up the least bad AND prevents the best, an intersection if you will... but maybe there's data to support the air out of balloon theory?

No offense, but now you're just being a bit patronizing. Is seeing the game well = seeing how you see it?

Yes, in my opinion. Not just me alone, of course. But of course I think I see the game well, and that's why I share my thoughts on it, that's why people solicit my opinion, etc. Looking back at my evaluation history, which I do meticulously, I have a good case. Not a perfect case, of course.

And I don't mean it to be patronizing, if that applies here, and I don't think my way is the only way...but beyond a reasonable doubt I think it's a way to create a positive impact/result. And there's a process to it. It's not random - like yearly save pct. leaders, for instance.

Perhaps your secret private analytics group will allow the public to see their methods one day then.

Well, that's what happens...that's the natural evolution of things, right? Like...save pct. was an "advanced stat" that was an analytic before...and then as teams realized it's lack of use or better methods were developed, that information becomes public. It no longer holds intrinsic proprietary value.

It's like when Corsi and all that noise came out...a lot of people here lapped that up and went "THIS IS THE WAY!" it was the golden plates...but that's because it was largely moved on from. That doesn't make it completely irrelevant and obsolete...but it needs context. It's not a substitute for proper evaluation.

So, we'll see over time where the goalie stuff goes. You're already starting to see things out there based on shot location and what not...and that's a piece of it.

Sure maybe all 5 saves were amazing (as improbable as that may be) and all 50 saves were the equivalent of stopping a beach ball (as improbable as that may be). I guess we can say that about every aspect of anything then, but we do sometimes use data/statistical probability to make determinations that we don't have time for our eyesight to catch, or even if we did, to freely remove our cognitive biases.

Well, you summed it up. The improbable situations - as they may be - will still require evaluation.

And yes, you once again nailed it...every aspect of anything is fair game. The game happens...stats get collated later.

And you're correct, sometimes we do use data to aid in the evaluation process. As one should...

 

[Dingo]

That's nice and all, but we still need to see the "Michael Farkas Certified Bad Goal Against That Really Lets the Air out of a Balloon" to really understand how bad Tim Thomas actually was and how SV% is a conspiracy

I genuinely like Mike and believe that he knows more about hockey than me. I don't think there is even an argument against that. But smart people can still be biased..... I feel like something actually happened between Mike and Thomas. He is a scout, he has probably met him. I'm not even saying this as a joke.

 

[Dingo]

I would as soon try to rate every baseball batter's worst, most untimely sucker strikeouts and their clutch, unbelievable hits as break down saves like this. Its not math and seems fraught with openness to inherent bias.

However, team defence, and stats vs backup + who that backup is don't have a comparison in baseball, where it is really just pitcher vs batter... thats more like shootout save% (only one aspect of stopping a puck, though)

for the record, i dont think Thomas is some slam dunk HOFr or anything. Just that he is definitely in the convo for top 3 American goalie since...
well kinda forever, actually. Like, he has to be..... top 10, no?

 

[Dingo]

I genuinely like Mike and believe that he knows more about hockey than me. I don't think there is even an argument against that. But smart people can still be biased..... I feel like something actually happened between Mike and Thomas. He is a scout, he has probably met him. I'm not even saying this as a joke.

hmmm, I've got my best guess

Telephone call, December 2004

GM "Hey Mike, how is Helsinki? Good, good. How's that Thomas guy that we sent you there to watch looking in net?"


MF (who hasnt watched him play and is hungover, under a pile of prostitutes... brain scrambling.... Eh, he is already way too old..... can't be good) "oh ya, Ive watched his last 9 games, boss.... i uh, gaurantee he will never amount to anything"

and, then the career goes downhill, along with the reputation, the days of Finnish ladies and vodka come to an end..... and then the hurting starts, and then the obsession. The poster of Thomas on the wall by the bed with the darts in it.... the nightmares about Thomas and the Stanley Cup, offering Mike a drink of Absolut......

Good Lord, I can see it all so clearly.

 

[Michael Farkas]

I genuinely like Mike and believe that he knows more about hockey than me. I don't think there is even an argument against that. But smart people can still be biased..... I feel like something actually happened between Mike and Thomas. He is a scout, he has probably met him. I'm not even saying this as a joke.

I can assure that that's not the case haha. It's not an interesting story at all, I've never met him...I don't think.

It's not him. If Martin Brodeur didn't exist (or whoever it was in his way) it would be Cechmanek. The thing about this particular case is that it's everything wrong* coming together at the perfect time.

How do I say this with a degree of intentionality...

Uhhh...it's a combination of so many glaring red flags to me, that gets through a threshold of fanfare, that makes me shake my head in disbelief haha

I'm legitimately surprised that I'm in the minority on this...full well knowing the stats...because no one goes to bat for Elliott...no one goes to bat for Cechmanek...so I know folks can see it...and I'm really surprised that this guy is different because I think he's an even more glaring case than Elliott is...and I hate Brian Elliott's game so much...

I gotta head in, but I have more on this if this is still going on later haha

 

[jigglysquishy]

I'm legitimately surprised that I'm in the minority on this...full well knowing the stats...because no one goes to bat for Elliott...no one goes to bat for Cechmanek...so I know folks can see it...and I'm really surprised that this guy is different because I think he's an even more glaring case than Elliott is...and I hate Brian Elliott's game so much...

I don't think you're in the minority.

Thomas wears the crown in the Cechmanek/Elliott category.

Goalies can put up world-beater stats in great conditions. And since they're the only goalie on the team (well, only starter) the numbers pop out.

When Kunitz was a PPG with Crosby we can point to Crosby as the reason. The relationship is clear because (Crosby PPG>>Kunitz PPG). It's harder for Thomas because Claude Julien doesn't have a statline. And the defensive prowess of Chara and Bergeron doesn't have easy stats to point to.

Thomas played in front of a system that didn't try to prevent shots, but tried to push low-quality shots. It's not far off from a system Carolina played in the early 2010s in front of Cam Ward. I don't blame him for succeeding in that style of system (Lord knows I couldn't), but it is a crucial piece of context when you are looking at SV% as a stat.

It's the inverse Martin Brodeur method. He kept shots low(er), which deflated his SV%.

 

[WarriorofTime]

With your question. you highlight a point that I've seen from you and there's a fundamental difference in how we view the game. I don't view the game from the perspective of "stats...and in". I don't watch a goalie and go, "hmm...I wonder what his average is?"

I'm looking at his process. What is his process for making a save at the level he's at?

My position is that goaltending is INHERENTLY a rate position. It just is. Same as baseball hitters/pitchers. There's no way around it. You can use GAA, SV%, 'Bad Goal' Rate, High Danger Save %, GSAA.. there is just no way around it.

I have not seen anything that cuts against that.

This is the problem with scouts, they fall too in love with their own scouting reports that they'll ignore real documented performance too much. This cuts both ways, fyi. Is it a spectrum? Sure. Of course. Once we're talking at the actual highest level though, especially in a HISTORICAL sense and not a future projecting sense, then throw the scouting reports out the window and look at what was accomplished above all.

 

[JackSlater]

I didn't consider Thomas at all for the American team. I don't know how many American teams I'd need to make before he came up as he's almost the opposite of what I prefer in a goaltender. I don't care for Vanbiesbrouck either, though I'd have him before Thomas. Quick was the clear number one.

I also strongly expect that at some point a goaltender stat heavily influenced by shot quality data, among other things, will become widely available and drastically change how people view save percentage, but that's for another thread.

 

[Michael Farkas]

This is the problem with scouts, they fall too in love with their own scouting reports that they'll ignore real documented performance too much.

Is that a problem? Like, a real one that...exists?

Why do only a couple of 55ish game seasons count as the "real documented performance"? Why are somewhat abbreviated 34 and 36 year old seasons the real deal and the other 18ish years of his career are the...what...exception?

Thomas likely didn't opt-in to the 1992 draft, so I'll give him a pass there...but he was passed over in 1993. He was drafted in a round that doesn't exist anymore in 1994.

They cut him loose.

He signed an ATO with Houston (IHL)...no sale. He got dumped to the ECHL.

To his credit, he fought to keep his career going...put himself in World Championships discussions...

Couldn't play over Pat Jablonski, Parris Duffus (a couple times), Garth Snow, Mike Dunham...and usually if you're a guy that keeps going to these things, they throw you a bone...but...ya know...he's not good.

He finally got to be a U.S. backup by beating out a goalie that couldn't hack it in the ECHL...

So, he signs in Finland - not a top three league in the world, maybe not top 5 at the time.

Edmonton lost Curtis Joseph after 1998 and brought Thomas to training camp. He was cut - in place of 34 year old Bob Essensa, Mikhail Shtalenkov, Steve Passmore, and the redoubtable Tommy Salo...

He was the worst goalie on Hamilton (AHL)...he goes back to Finland.

1999 - training camp with Tampa. Can't beat out Cloutier, old Puppa, Zac Bierk, Kevin Hodson, Rich Parent, Dieter Kochan...cut. Sent to Detroit (IHL)...worst goalie on that team by the numbers.

Back to Europe...

Boston gives him a look in 2001...again though, Byron Dafoe, John Grahame, Andrew Raycroft, and Jeff Maund (who would quickly end up in the Italian League) beat him out...cut...and shipped back to Europe.

Back for more...can't get playing time over Shields, Hackett, Grahame as a 28 year old now...

Lockout.

So, now the goaltending landscape is really shaken up by the rule changes. Not a ton of goalies adapted well to this...so there's opportunities all over.

Thomas - now 31 - finally gets a sniff as young Toivonen loved letting pucks go through him and their anticipated bell cow, Andrew Raycroft, turned out to not be super adaptable to the new league...

We all know the NHL story. We've all seen the with Julien vs without Julien numbers and they're staggering, of course.

I think it's generally well known that he wasn't, ahem, an impressive goalie...he's a bad skater, he doesn't anticipate well, he can get beat from virtually anywhere...because he doesn't have a good process for making saves. And like Cechmanek and some of these fly-by-night guys, it's like facing a knuckleball pitcher a little bit...you can change your swing and do all that kind of stuff to figure it out, or you just hope to go 1 for 3 with a walk and move on to the next game.

"Yeah, well, who cares..."

Well, everyone that evaluates goaltenders. He only had a short window of success in a very particular situation and generally wasn't considered NHL caliber around that time. It's not like he was up and down and around like Barrasso, for instance...early success, then kicked around, then more success, then not so much, then one last ride in '98, and then retirement age stuff...

He wasn't able to stay in the top 6 of any organizations' goalie depth chart for many years with many teams that didn't have goaltending...

He also wasn't having success in the 2nd, 3rd, 4th best leagues in the world either...he had to go below that.

So even if you want to throw out the talent evaluation of it, if that's just not for you (royal you)...it still shouldn't pass the smell test.

 

[WarriorofTime]

Is that a problem? Like, a real one that...exists?

Why do only a couple of 55ish game seasons count as the "real documented performance"? Why are somewhat abbreviated 34 and 36 year old seasons the real deal and the other 18ish years of his career are the...what...exception?

Because I don't give a crap that you (the royal you, not necessarily you) feel proud of your scouting report that said "NOT AN NHL GOALTENDER".

That was proven wrong, performance is what matters when you look back retrospectively, not how pretty or technical or fluid we think someone's game was.

"Yeah, well, who cares..."

Well, everyone that evaluates goaltenders.

And there you go again, with your gatekeeping secret illumanti club that everyone should implicitly trust and never question. OK.

 

[Michael Farkas]

Ok, sure...

Now, back to the thread. We're building a team. We have a bunch of goalies available that - forget technical, we're not interested in the game itself at an atomic level - had success in various situations, across multiple eras, multiple teams, over more time, etc. etc.

Then we have a guy that had a very small window of success in one particular situation and no others.

How could one justify adding the latter to a team at this level?

 

[WarriorofTime]

I also strongly expect that at some point a goaltender stat heavily influenced by shot quality data, among other things, will become widely available and drastically change how people view save percentage, but that's for another thread.

GSAA/GSAE is already becoming more and more the "default" stat

 

[JackSlater]

GSAA/GSAE is already becoming more and more the "default" stat

It's definitely a step in the right direction. I'm not even sure exactly what stat I would want or saying that save percentage is trash. It's just not close to being an argument ended 99% of the time. I'm sure that I'd prefer Quick, or Hellebuyck, or Miller, or Barrasso, or even guys like Richter, Vanbiesbrouck, and Bishop to provide steadier and more predictable goaltending than Thomas could in a three team tournament where he plays behind probably the weakest team defensively. I don't expect most to agree but I'm pretty firm when it comes to Thomas and what I saw... sort of a homeless man's Hasek who found his way onto a team with great defensive personnel and strategy.

 

[Regal]

Thomas was a very good goalie at his peak, but he was highly aggressive and would often give up big rebounds and move around wildly in the crease. The Bruins did an excellent job at limiting the ability for team to get to those rebounds or make cross crease passes to exploit his aggressive movements. I don’t think many goalies necessarily have the same success in his position but I’m also not sure if he has the same success in other situations. In a best on best I think it’d make sense to have a less erratic goalie (unless it’s Hasek), particularly if he’s going to be facing Gretzky and Lemieux.