2005-06 NHL Goal Differentials

[Clumsyhab]

Amazing stats!!! Theodore was the least efficient goalie last year lol. Hopefully, he'll do better next year with the Avs, we still like him in Montreal.

 

[Doctor No]

One item that I would like to put forth is that, save % is GREATLY skewed by the team in front of the goaltender. A team like Minnesota is structured to allow lower quality shots from farther out than is a team like Ottawa. So, save percentage and all of its deviations are skewed, IMO.

I concur that save percentage is not perfect; as I mention in post #17, what I'd like to do is risk-adjust each individual shot on goal for its level of difficulty.

What can be shown, however, is that save percentage is much less team-dependant than either goals-against average or wins (in other words, if Goaltender X is a free agent yet to sign, it's easier to predict his 2006-07 save percentage than either of the other two statistics).

The two things this method does is (1) normal save percentage across eras, and (2) leverage the effects of save percentage based on playing time. As I mentioned in the original post, if you're not a fan of save percentage, this probably won't be your thing.

 

[Doctor No]

On the other hand, I've never heard a Grant Fuhr supporter use his regular-season save percentage as evidence for his greatness, so it's not too surprising that this metric does not fare him well.

Grant Fuhr supporters generally use one of these two arguments:

1. Because of the style of play Fuhr's teams exhibited, Fuhr faced more difficult shots on goal than a typical NHL goaltender of the 1980s. His traditional statistics are therefore penalized.
2. Fuhr played better in clutch situations than a typical NHL goaltender of the 1980s, an ability which is not reflected in his statistics.

I happen to believe point #1. Warning - the remainder of this paragraph is my opinion only, and not based on any mathematics of any kind. If your team is in the other team's zone all day long, what kind of shots on goal are you going to face as a goaltender? Breakaways. Two-on-ones. Three-on-ones. I'm not suggesting that every single shot Fuhr faced was a breakaway - far from it - but I'm saying that it's reasonable to assume that he saw more tough shots than a typical NHL goaltender.

Point #2 I'm not a fan of. To my knowledge, it's never been looked at for goaltenders, but a lot of work has been done trying to find "clutch ability" in other sports. What has been found is that any "clutch ability" is washed out by the random noise.

What could be done is going through, game by game, and looking at situations which could be considered "clutch" to see if Fuhr performed any better than an average NHL netminder. I have not done this - I don't have anywhere near the resources available to do a task of this magnitude.

One thing that could be done is to look at Fuhr's game log on my website:
http://hockeygoalies.org/bio/fuhr.html
and see if he performed better in "big" games. You'd have to fairly define what you mean by "big" games, but it could be interesting to see if Fuhr did better than, say, Patrick Roy (http://hockeygoalies.org/bio/roy.html).

As a way to end this post, I think we can all agree that playoff games are "big games". How did Fuhr do in the Stanley Cup playoffs?

Fuhr recorded an even 90.0% save percentage in his Stanley Cup career, with a league-average weighted save percentage of 89.5%. That gives him a goal differential of 17.63, meaning that Fuhr was almost 18 goals above average over his playoff career.

If Fuhr had played as well in the regular season as he did in the playoffs (this is the same calculation I did a few posts back with Hasek and Emery in the 2006 playoffs), he would have been 95.33 goals above average. As mentioned above, Fuhr was actually 37.65 goals below average over his NHL career, a difference of 133 goals. This is the best evidence that I know of that Fuhr was better in the big games.

 

[Doctor No]

Amazing stats!!! Theodore was the least efficient goalie last year lol. Hopefully, he'll do better next year with the Avs, we still like him in Montreal.

As an Avalanche fan, I'd certainly love to see it!

For his career, Theo is currently 38.92 goals above average, placing him 32nd on the leader board between 1982-83 and now. Of course, he's still pretty young so he has plenty of room to climb the charts.

 

[Doctor No]

What a surprise, Thibault and Caron were in the very bottom of the list. But who'd have guessed that Caron would actually rate worse than Thi-bound?

Part of that's because Caron played 500 more minutes than Thibault.

Just as an above-average goaltender is more valuable if he plays more, a below-average goaltender is less valuable if he plays more.

Also, a "below average" goaltender still has value, so a better metric may be Goals Above Replacement-Level, meaning how many goals better is a particular netminder than someone available on the waiver wire.

Although I'm still fiddling with the exact location of "replacement level", Caron does slightly better than Thibault in this metric.

 

[barfy2000]

I find it interesting to see Roberto Luongo so high on the career list already. Really says a lot about him I guess.

 

[handtrick]

I think this stat really gives Vokoun the credit he deserves, but seldom receives.

 

[JussiM]

I think this stat really gives Vokoun the credit he deserves, but seldom receives.

Damn right!

Amazing work on the stats, well done!

 

[Cawz]

Man, this really shows what a gong show Edmonton's goaltending was before Roloson.

And really shows what gong show the NW will be next season. #1, 3 and 6 from the NW.

 

[SENSible1*]

How much did Hasek's injury cost the Senators? If we extrapolate his regular season statistics to the 2006 Ottawa playoffs, he would be worth 6.70 goals above average. Ray Emery totalled 1.76 goals below average.

That's about 0.85 goals per game Hasek would have stopped over Emery (under these assumptions, of course). Would that have made a difference in the Buffalo series?

Great work on the stats. Thanks for posting them.

Hasek's injury cost the Sens the cup. It's nice to have a statistical basis for the theory, but in this case not really necessary.

 

[Squeaky]

I appreciate it when someone puts that much effort into this kind of work.

One item that I would like to put forth is that, save % is GREATLY skewed by the team in front of the goaltender. A team like Minnesota is structured to allow lower quality shots from farther out than is a team like Ottawa. So, save percentage and all of its deviations are skewed, IMO.

My firm belief is that there is no goaltending statistic that can be relied on to compare goaltenders. The team in front of the goaltender is a major part of how the numbers end up.

As I have said before, put Grant Fuhr on the Canadiens in the 80s and he probably wins 3 Vezinas and has 50 career shutouts.

I agree that goalie stats aren't usually the most terribly useful ones around, but there is still great value, especially if you look at them right. Take Montreal for last year as an example. They played most of the year under Theodore, who scored a league second-worst -20.81 then went straight to Huet, who scored fifth-best at 30.11. For the habs, this is a very telling stat.

Here's Hoping Huet stays that good, and Theodore that bad

 

[fastb]

Great work on the stats. Thanks for posting them.

Hasek's injury cost the Sens the cup. It's nice to have a statistical basis for the theory, but in this case not really necessary.

That seems a bit of a stretch to me. Not that I think they couldn't have won the Cup with Hasek, but you can't conclude that with the addition of Hasek everything else would have been equal. I think the most you can conclude is that his absence may have cost them the Buffalo series, but anything after that is just speculation.

 

[Cawz]

That seems a bit of a stretch to me. Not that I think they couldn't have won the Cup with Hasek, but you can't conclude that with the addition of Hasek everything else would have been equal. I think the most you can conclude is that his absence may have cost them the Buffalo series, but anything after that is just speculation.

Nope, its a fact. The Hasek injury lost the Sens the cup, the Roloson injury lost the Oilers the cup, the injuries to all their defensemen lost the Sabres the cup...

Didnt you know that no team actually wins the cup - all the other teams just lose it.

 

[SPARTAKUS*]

I agree that goalie stats aren't usually the most terribly useful ones around, but there is still great value, especially if you look at them right. Take Montreal for last year as an example. They played most of the year under Theodore, who scored a league second-worst -20.81 then went straight to Huet, who scored fifth-best at 30.11. For the habs, this is a very telling stat.

Here's Hoping Huet stays that good, and Theodore that bad

Theo is to damn good to have the kind of season he has last season. This guy is a true competitor and will have a huge season for the Avs this season. I predict that he will be a Vezina nominee.

As for Huet well fans in Montreal see him as the second coming of Christ. And he only had one good season. And what did he do in the playoffs nothing he had a 2-0 series lead and got beat 4 games in a row. Personally I think he will be a one year wonder. He's a good goaltender not a great goaltender. He's not in Theo's league.

 

[optimus2861]

Bumping after link from another thread...

Goalies World magazine recently computed the save percentages of the top ten goaltenders in every season clear back to 1954-1955 (issue number 55, cover date December-January 2006) by amassing the individual game results. While they're not official stats, it should make a limited "goal differential" analysis possible, at least for the goalies listed.

Jaw-dropper: Jacques Plante, 1970-71 Maple Leafs, at the age of 41: 0.942 save percentage! The next highest? Plante, again, in 1968-69, still no spring chicken at 39, at 0.940.

 

[Doctor No]

Jaw-dropper: Jacques Plante, 1970-71 Maple Leafs, at the age of 41: 0.942 save percentage! The next highest? Plante, again, in 1968-69, still no spring chicken at 39, at 0.940.

Both of those are definitely eye-popping!

My understanding (I haven't read this issue of GW in a few months) is that their research comes from Edward Yuen, who has been putting these together for well over a decade now, and he has been very diligent in his error checking. In other words, his results are probably going to be as good as we'll ever get (barring a time travel device of some sort).

Klein and Reif's Hockey Compendium printed the seasons available at the time, including all of the playoff totals since 1953 (partial 1952). I've posted the top playoff goal differentials here:

http://hfboards.com/showpost.php?p=6223972&postcount=36
http://hfboards.com/showpost.php?p=6223972&postcount=38
http://hfboards.com/showpost.php?p=6223972&postcount=41

and am working on a summary of each season for my goaltender website.

I would love it if Goalies' World would release all of the year-by-year statistics; Yuen might himself, although he's hard to get ahold of (I tried once).

Thanks!

 

[Whatever Man*]

Nice Job, Dr No.

Putting that math degree to good use, I see.

Goals per year difference. Very interesting.

Hasek's run of over 54 goals per year is incredible.

 

[Doctor No]

Thanks! Yeah, Hasek's five-year run at the top was phenomenal.

I can't wait to get ahold of Edward Yuen's older save percentages to see how some of the older (Dryden, Plante, Parent, Hall, Bower, Esposito) goaltenders fare.

 

[BM67]

Looking at the stats in the Hockey Conpendium, I find it quite remarkable that Plante had a sv% of .900 or better for all 12 seasons covered and 14 of his 16 playoffs, with only his last two, 72 & 73, falling below .900.

 

[Doctor No]

I was surprised as well - especially with how well Plante performed at the later ages of his career.

 

[Rand]

I happen to believe point #1. Warning - the remainder of this paragraph is my opinion only, and not based on any mathematics of any kind. If your team is in the other team's zone all day long, what kind of shots on goal are you going to face as a goaltender? Breakaways. Two-on-ones. Three-on-ones. I'm not suggesting that every single shot Fuhr faced was a breakaway - far from it - but I'm saying that it's reasonable to assume that he saw more tough shots than a typical NHL goaltender.

Plausible idea, but if so why did his backups always post similar and indeed usually better SV % then Fuhr himself did?
They were behind the same team so it seems logical enough to me that they would generally face a similar quality of shots and hence be penialized to the same extent in any pure measure of SV %.

 

[Lowetide]

Plausible idea, but if so why did his backups always post similar and indeed usually better SV % then Fuhr himself did?
They were behind the same team so it seems logical enough to me that they would generally face a similar quality of shots and hence be penialized to the same extent in any pure measure of SV %.

I've wondered the same thing. Billy Smith has a similar problem in his record, with Resch and Melanson basically marching in lock step with his numbers while backing up Smith.

It's a tough question, and I don't have the answer.

 

[Doctor No]

Plausible idea, but if so why did his backups always post similar and indeed usually better SV % then Fuhr himself did?
They were behind the same team so it seems logical enough to me that they would generally face a similar quality of shots and hence be penialized to the same extent in any pure measure of SV %.

I'll give it a stab.

In a two-goalie situation, the lesser goaltender should be facing poorer-quality opponents. And if he's not, then the coach isn't maximizing his team's chances of winning.

 

[jstreet]

I love seeing Vokoun near the top of that list!

 

[misterchainsaw]

Luongo FLO 32.62
Auld VAN 2.32

Hmmmmm.....