View Single Post
Old
02-08-2013, 07:47 PM
  #75
Canadiens1958
Registered User
 
Canadiens1958's Avatar
 
Join Date: Nov 2007
Posts: 13,818
vCash: 500
Assumptions and Data

Quote:
Originally Posted by ContrarianGoaltender View Post
I guess it could, theoretically. In the real world, though, it did not.

I'll echo the others who say that the suggestion that Dominik Hasek inflated his save percentage from facing more rebound chances is completely evidence-free, and a major example of the problems of rating goalies subjectively based on their style rather than their results.

First of all, if Hasek created that many extra shots on a regular basis, he would have faced far more shots against than the other goalies he played with. This was not the case.

1990-91: Hasek 28.6, Backups 27.3
1991-92: Hasek 24.4, Backups 25.0
1992-93: Hasek 30.2, Backups 32.0
1993-94: Hasek 27.7, Backups 31.5
1994-95: Hasek 30.3, Backups 28.4
1995-96: Hasek 35.3, Backups 34.7
1996-97: Hasek 32.4, Backups 34.5
1997-98: Hasek 30.6, Backups 31.2
1998-99: Hasek 29.5, Backups 29.2
1999-00: Hasek 27.2, Backups 26.1
2000-01: Hasek 26.5, Backups 27.2
2001-02: Hasek 25.6, Backups 27.0
2003-04: Hasek 23.8, Backups 26.3
2005-06: Hasek 27.9, Backups 28.8
2006-07: Hasek 23.5, Backups 26.0
2007-08: Hasek 21.8, Backups 24.7

The overall average shots faced per 60 minutes are 28.3 for Hasek and 28.1 for his backups. If you weight the backup numbers by Hasek's minutes played in each season, their combined number ends up being 29.7, or 1.4 more than Hasek per 60 minutes. Either way, Hasek did not face a lot of additional shots compared to his teammates, which completely defeats the claim being made about him regularly creating 3 or 4 additional chances at a pop.

Secondly, shots on rebounds are extraordinarily dangerous scoring chances. Read this article for a sample of rebound shot data that estimates the shooting percentage of rebound shots in the NHL at around 27%. Hasek's career save percentage was .922, which means that for rebound saves to help his save percentage, opposing shooters would have had to shoot 7.7% or worse on rebound shots. Is it plausible that Hasek was that far ahead of average over his entire career on those types of chances? In my view, even though Hasek was surely the best ever at dealing with those chances, the observed margins in goaltending do not come close to supporting anyone being that far ahead of everyone else.

Thirdly, the number of shots on goal from rebounds in the NHL is actually relatively rare, around 3 for both teams combined per game in the samples I've seen. It can depend a bit on the exact definition of a rebound shot, but it's probably unlikely that Hasek was facing more than 4 rebound shots against in the entire game, much less in the course of one play. Most of the time the defensive team clears the rebound or ties up the opposing team's sticks so they don't get a shot off. And that makes sense, because even with a great goalie in net rebound shots are very likely to go in the net.

From his list and from the discussions, I don't think C1958 was necessarily inconsistent with his rankings, but I do strongly disagree with the relative rankings of different goalie skill sets that informed his votes, because I don't see them as being supportable by the statistical evidence. By far the most important part of goaltending is making the first save, and when you do it as well as Hasek did it (Taco McArthur has him at 80-90 goals above replacement level at his peak), it becomes hard to justify claiming that other skills make up the gap.
Your presentation is based on a number of assumptions and pulls data from different eras.

Hasek played in the NHL from 1990 thru the dead puck era thru 2007-08. Your rebound data is from the 2007-08 era and does not reflect how the game was played pre the 2004-05 lockout. Rules changed significantly post lockout. The 2 second allowance is very generous for rebound goal since it covers the complete offensive zone as opposed to the slot.

You are operating under three assumptions that have not been vetted or supported.

1.) that the other goalies on Hasek's team were better than him in terms of rebound control. If you look at his teammates, only two were within 30 spots of him in this project - Belfour and Fuhr. No claim has been made that Belfour and/or Fuhr were better than Hasek in terms of rebound control. In Buffalo, after Fuhr was traded Hasek's back-ups were rookies - Biron or career minor league veterans - Shields, Roloson. Weaker rebound control in all instances.

2.) assuming that the data vs 1990 back-ups has meaning in terms of comparisons against other goalies vying for the top 40 spots and who retired before 1990. The only meaning it may have is in comparison to Roy and Brodeur. But you did not produce the numbers for Roy and Brodeur. Compare Hasek's numbers vs Brodeur and you will see the rebound control difference.Then watch game films to compare the statistical with the visual. Brodeur is the only goalie from the post 1990 era whose rebound control would compare with the pre 1990 goalies.

3.) your data would be interesting if it factored in the opposition. Are you assuming that against Hasek the opposition had more or fewer shoots, than against the rest of the league teams.

Interesting but far from complete.

Canadiens1958 is offline   Reply With Quote