>>>As I said above, I've never seen Herlofsky play, so I can't pass final
>>>judgement on him. I can, however, tell you why his Perserverence Rating
>>>is so low. Basically, while he has a great GAA (3.14), his save% is very
>>>ordinary (84.72%). All the other goaltenders mentioned in the top 9 have
>>>save% from 87-90%. He also faces an average of only 20 shots per game,
>>>while most of the others listed are facing 29-33 shots per game.
>>
>>So goalies with very good team defense in front of them who face fewer shots
>>per game automatically suffer using this "Perserverence Rating". Is that
fair?
>
>Well, the rating is designed to measure a goaltender's perserverance under
>pressure.
>
>A goaltender who faces fewer shots per game will not automatically
>suffer. Blair Allison of Maine is only seeing 23 shots per game, but his
>Perseverence Rating is 930.96.
Not automatically, no, but this 'rating' does favor goalies who face
a higher number of shots per goal. Besides, Allison's rating just puts him
into the good category ahead of the mediocre category. I am no Maine fan,
but Allison is the best goalie in Hockey East this year and he is better
than 'good'. Again, the low shots per game total he faces hurts him in this
'rating'.
>>Again, as Mike Machnik asked, where did the "rating" come from? How was the
>>range of great to horrible decided? What else was this "rating" measured
>>against?
>
>The rating was developed by Jeff Klein and Karl-Eric Reif. It was published
>in their book "The Klein and Reif Hockey Compendium" (McClelland and Stewart,
>1987). This statistic is explained fully in Chapter 8 of the book. The main
>rationale for this new statistic is that save percentage does nothing to
>measure a goaltender's workload. The assumption, I guess, is that a goalie
>who only faces 20 shots per game has an easier time of it than one who faces
>40. It was developed based on NHL statistics from the early-to-mid 80s.
>The ranges of what is good, and what isn't were developed, I would assume,
>based on those. They may vary in the NC$$, but I don't think they vary all
>that much. 881.44 is still pretty bad.
I have never heard of these to authors or their book. I have also
never seen it mentioned by anyone anywhere since they published their book
in 1987. I would have to say the the rest of the hockey world has dismissed
their 'rating' as an unreliable way to rate a goalie.
The assumption, you guess? Again, where is the actual reason for
this 'rating'? There must be a solid foundation for this 'rating' and it
must be explained. Otherwise it is meaningless.
If it was developed with several years of statistics used, than what
is the minimum number of games that it can be reliable? Nine? Twenty?
Fifty? One hundred? Two hundred?
Why would it vary in the NCAA?
>>Is it fair that BU's goalies suffer because they have good team
>>defense in front of them (most games)? BU has allowed an average of only 23
>>shots per game, while they average taking 37 shots per game, so according to
>>this rating they are bad goalies despits having good GAA's, average saves
>>percentages and winning records.
>
>First, it is impossible to completely divorce a goaltender's performance from
>that of his team, but I think this comes as close as I've seen to doing that.
You may, I do not. I can look at a goalie's GAA, saves percentage
and win-loss record and tell if he is a good goalie. I do not need an
unsupported 'rating' to decide.
>>I did quick calculations for Tom Noble and Derek yesterday and I came up
>>with a 905 overall rating for Derek ("very poor") and 918 for Tom ("poor"),
>>and a 912 for BU gaoltending overall ("poor"). Then I did the calculations
>>changing only the shots per game (i.e., the saves percentage remains the
>>same). If Tom and Derek faced an average of 37 shots per game instead of
>>the 24 (Tom) and 22 (Derek) they do face guess what? Tom becomes becomes
>>"very good" (940) and Derek becomes "good" (931).
>>
>>I think that a supposed goalie rating that can change so much based on a
>>factor beyond the goalie's ability to control (shots per game faced) is a
>>very unfair and unreliable.
>
>But two paragraphs ago you were saying that Herlofsky and Noble were good
>because they had good win-loss records and good GAAs. There is no goalie
>stat that is more out of his control than win-loss record. Herlofsky's
>GAA is also a function of his defence. Since he's only seeing 20 shots
>per game, I would be worried if his GAA were much more than 3.14.
They are good goalies. I was trying to show that this 'rating' is
fatally flawed. A goalie's rating can change based on shots faced. Just
because a goalie faces more shots does not automatically make him better,
but this 'rating' does make him better.
What about two goalies with identical saves percentage and number of
games played: Goalie A has a 88.5% saves percentage, has played in 50 games
and has faced 40 shots per game. Goalie B has a 88.5% saves percentage, has
played in 50 games and has faced only 20 shots per game. Goalie A has a
'rating' of 951.67, an excellent rating. Goalie B has a 918.33 'rating' a
poor 'rating' . Do you still say Goalie A is better? The only reason
Goalie A has a higher 'rating' is because he faces more shots per game.
Again, is that far? The goalie can not control shots faced per
game, yet this 'rating' makes it a large part of its equation.
>Just for fun, I did some number-playing as well. It turns out that if
>Herlofsky's save% were 0.8862 (the average of the other 8 HE goalies),
>his Perseverence Rating would go up to 920.52. That part of it is within
>his control.
Yes, but since he had a very poor start it will be hard for him to
reach an 88.6 saves percentage, especially after Friday night's game. Derek
has already played in 19 games he will only play in 25 or so games total
this year, so he will be hard pressed to improve his statistics greatly for
the year.
>And I really didn't intend to end up dumping on BU goalies. I have nothing
>to do with Maine, honest. :-)
I never said you did. Besides, this 'rating' also dumps on Blair
Allision.
>But it does look to me that Herlofsky is overrated, at least this year.
>I would be interested in seeing his stats for last year. If I find them
>in the archives, I'll try to figure out his P.R. for last year.
No Derek is not overrated, but he did play poorly at the start of
the year. Of course, having his friend and goalie buddy J.P. McKersie
seriously injured and unable to play this year did affect him mentally at
the beginning of the season.
As for Derek's BU career, his 'ratings' would be:
Saves% SPG "rating"
1994-95 .862 22.39 899.31
1993-94 .909 27.23 954.38
1992-93 .893 26.43 937.05
1991-92 .912 27.94 958.57
Career .893 25.73 935.88
>--
>John C.K. Edwards Stats Geek, Ottawa Jr Senators (23-22-3 51pts) (CJHL)
>Carleton U., Law IV I don't give a damn about being liked,
>[log in to unmask] but I sure as hell intend to be respected.
>All Canadian [Tier II] Junior A standings: http://chat.carleton.ca/~jedwards
Sean Pickett
Go BU Terriers, 1994 Hockey East Champions!
BU Hockey Page: http://www.tiac.net/users/spickett/hockey.html
[log in to unmask]
|
