2005-06 NHL Goal Differentials

[bfire]

Nice work

 

[Injektilo]

Again, this is good stuff, thanks for putting it together.


If possible, you might wanna stick to even strength SV% in the future, as goalies that have to kill alot of penalites are obviously going to be on the ice for more high percentage scoring chances. If you can stick to ESSV%, that might clear up a bit of the "quality of shots" mess that comes with SV%.

 

[Doctor No]

Thanks! You're right about ESSV%, although even-strength save percentage can be quite hard to come by, especially for any historical purposes.

The ultimate goal is a risk-adjustment of every shot taken in a game, so that (for instance) a 60-foot slapshot can be compared with a breakaway backhander from five feet out in terms of relative difficulty.

This has the obvious problems, but the results will be a lot more illuminating.

 

[NYR94]

Nice job. It's good to see guys like Luongo and Vokoun near the top, where they're supposed to be.

 

[seventieslord]

Top twenty careers (1982-83 to present):

Note that save percentage was not made official until 1982-83; hence, seasons prior to this are not considered.

Extra note - the difference between Roy/Hasek and the rest of the goaltenders is phenomenal.

[font=courier new]Patrick Roy               429.91
Dominik Hasek             392.49
John Vanbiesbrouck        192.94
Curtis Joseph             188.58
Ed Belfour                165.20
Martin Brodeur            152.61
Roberto Luongo            133.69
Kelly Hrudey              125.13
Andy Moog                 111.89
Guy Hebert                108.85
Mike Richter               92.17
Tom Barrasso               87.89
Bob Froese                 83.50
Sean Burke                 80.63
Ron Hextall                77.19
Pete Peeters               76.72
Rejean Lemelin             76.57
Glen Hanlon                73.02
Billy Smith                71.42
Daren Puppa                71.38
[/font]

Interesting! This looks about right to me, although I would think Brodeur would have found his way into thine above the Beezer by now. Maybe I'm overrating him too - maybe his solid defensive team and lack of shots against help him even more than is commonly believed.

If you had asked me to rank these goalies on how I thought they'd turn out, I also would have thought Barrasso would take 4th spot.

So correct me if I'm wrong, but for this to be evened out for games played, I could just divide by games played, right? Hasek would have the per-game edge on Roy, but not by much.

Is there enough data to do this for playoffs?

 

[Doctor No]

So correct me if I'm wrong, but for this to be evened out for games played, I could just divide by games played, right? Hasek would have the per-game edge on Roy, but not by much.

That would work, although to account for partial games played, I'd multiply by 60 and divide by the number of minutes played.

As for Brodeur, I'd agree that I think he would be higher. Note that this method doesn't take into account any shots Brodeur prevents by his stickhandling (which many people believe to be non-trivial).

For playoffs, I have them posted in another thread - I'll post the link shortly (after a quick trip to the search feature).

(Edit: http://hfboards.com/showpost.php?p=6235173&postcount=41 has links to the postseason totals.)

 

[Hockey Outsider]

I'm not sure if this has been posted yet...

Here's the Goal Differential data for the Original Six era, and a few other seasons. These are based on the work of Edward Yuen, who manually calculated shot/save data for every goalie in these years (see "Hockey Compendium").

Year-by-Year Leaders
1955 Harry Lumley -- 31.3
1956 Jacques Plante -- 23.9
1957 Glenn Hall -- 38.7
1958 Gump Worsley -- 26.4
1959 Jacques Plante -- 43.3
1960 Johnny Bower -- 28.6
1961 Johnny Bower -- 31.3
1962 Jacques Plante -- 43.2
1963 Gump Worsley -- 19.1
1964 Glenn Hall -- 32.2
1965 Glenn Hall -- 20.3
1966 Johnny Bower -- 32.4
1967 Edward Giacomin -- 24.8
1971 Jacques Plante -- 52.1
1975 Rogie Vachon -- 60.9
1976 Ken Dryden -- 65.1

Observations:
- I was really surprised that Terry Sawchuk never led the league. In fact, he finished second only once and third once. That's a pretty poor performance for somebody generally regarded as one of the best goalies ever. Still, it's worth considering that Sawchuk was an elite playoff performer (which naturally is more important than regular season performance). Additionally, many of Sawchuk's best years occured before '55, so its possible that Sawchuk had some great years according to this method, but we just haven't uncovered the data yet.
- Plante was the most dominant goalie by a small margin; he led the league in GD 4 times and was runner-up 4 times. Hall and Bower each led the league 3 times and were runner-up 4 times.
- Ron Low, Al Rollins, Marcel Paille and Eddie Johnston were the worst goalie in the league twice each.

Highest Peak Value (1955-1967 only... average of three best seasons)
Jacques Plante 36.8
Glenn Hall 33.6
Johnny Bower 31.7
Gump Worsley 21.3
Terry Sawchuk 15.8
Denis Dejordy 13.0
Harry Lumley 12.4
Charlie Hodge 6.3
Don Simmons 6.2
Bruce Gamble 1.9

Pretty big drop-off in top-end talent, eh? This confirms that Plante, Hall and Bower were the most dominant goalies of the Original Six era (according to GD, anyway)

 

[Doctor No]

I'd love to see Yuen on these boards, for what it's worth, mainly because I'd love to know how his data is organized. (But also because I think he has a wealth of information about goaltending history)

Specifically, it may save me a lot of time in compiling game logs (e.g., http://hockeygoalies.org/bio/roy.html at the bottom of the page) for these seasons.

 

[Hockey Outsider]

Total Career Value, 1955-67
Glenn Hall 218.1
Johnny Bower 196.8
Jacques Plante 164.0
Gump Worsley 67.4
Denis Dejordy 36.8
Charlie Hodge 14.8
Rogie Vachon 5.2
Gerry McNamara 1.7
Ross Wilson 1.4
Gary Bauman 0.3

- Again, this confirms that Hall, Bower and Plante were by far the best three goalies from 1955-67. It also confirms the rapid drop-off in talent after those three netminders.
- Simmons, Paille, Chadwick, Gamble and Rollins all hover at around -60 and Eddie Johnston is at -100.4. Were they that bad in comparison to Plante/Hall/Bower, or were there no other decent goalies avaialbe?
- Sawchuk finishes at 0.0: he was exactly average over that span. Again, this understates Sawchuk's true value, because this data excludes the playoffs, and it also fails to cover Sawchuk's best years, '51-'54.
- Data for the 70's is limited, but the big names are who you'd expect. Dryden finishes 5th and 1st in his two years as a starter; Esposito is 2nd, 3rd and 4th in three years; Giacomin has years finishing 1st and 3rd (but also a few lousy years); Parent, Vachon and Smith look good, but the data for them is quite limited.

I should mention my calculations are slightly different than Dr. No's.
- I remove the goalie's shots and saves when calculating the league average. In a six-team league, a single goalie can have a major impact on the average.
- Instead of using actual shots faced, I used "expected" shots faced. Expected shots faced = minutes played * league average shots per minute. I thought it was unfair for goalies to have higher/lower scores because they play in front of defenses that allow more/fewer shots than average.

 

[Doctor No]

I should mention my calculations are slightly different than Dr. No's.
- I remove the goalie's shots and saves when calculating the league average. In a six-team league, a single goalie can have a major impact on the average.
- Instead of using actual shots faced, I used "expected" shots faced. Expected shots faced = minutes played * league average shots per minute. I thought it was unfair for goalies to have higher/lower scores because they play in front of defenses that allow more/fewer shots than average.

I debated both of these ideas in my mind (and am continuing to debate them). Especially after calculating the Original Six playoffs (in the other thread), I'm leaning towards changing the first to Hockey Outsider's method.

 

[Over the volcano]

Great stuff Doc!!