Overworking a goalie is a myth (with statistical analysis
Basically I've gotten sick of everyone in the hockey world using the same old excuse every time a goalie plays a lot of games and has poor stats and/or a poor post-season - he's tired and over-worked. I couldn't find any analysis done on this topic online so I spent a few minutes and quickly did one myself.
All my samples were of goalies that played at least 68 games 2+ years in a row and were sandwiched by 2 years of 63 games or less. I have included all the samples past 1990. I would add data from before 1990, but I can't find accurate data for years before then. In spite of that, there likely isn't a great data more samples before 1990 due to goalie usages in that era.
Here are my two tables of results. The second includes goalies who've played only "before" years (Kiprusoff and Lundqvist) due to a lack of samples who've played 3+ years of 68+ games. Marty Brodeur is excluded from the first table due to him playing 10 years of 68+ games and having too much of an impact on the results.
DURING - all stats from years of playing 2+ consecutive years of 68+ games
BEFORE - stats from year prior to playing first year of 68+ games
AFTER - stats from year after playing last year of 68+ games
EXAMPLE: Guy Hebert's relevant years:
Year Team GP
1997-1998 ANA 46
1998-1999 ANA 69
1999-2000 ANA 68
2000-2001 ANA/NYR 54
DURING = stats from 1998-99 and 1999-2000 (years where he played 68+ games)
BEFORE = stats from 1997-98
AFTER = stats from 2000-01
2 Years of 68+ Games (Excluding Brodeur) – Only Goalies with Before and After years
TOTAL GP MIN/60 GA GAA SA SV SV% DURING 1861 1801.37 4771 2.65 53185 48414 0.910 BEFORE 539 512.97 1474 2.87 15606 14132 0.906 AFTER 633 610.3 1579 2.59 17225 15646 0.908
3 Years of 68+ Games (Including Brodeur)
TOTAL GP MIN/60 GA GAA SA SV SV% DURING1969 2081.32 4951 2.38 56996 52045 0.913 BEFORE 338 373.25 903 2.42 10657 9754 0.915 AFTER 216 272.7 655 2.40 7726 7071 0.915
It's pretty clear looking at the differences in GAA and Save% that there is no discernable effect of playing 68+ games several years in a row. GAA rises and Save% improves for 3+ years, but both are by negligible amounts. For 2+ years, GAA improves, but Save% decreases. Again, both are by negligible amounts. The only stat that has any improvement worth noting shutout% (which is simply SO/GP).
Last edited by JayP; 04-29-2009 at 10:52 PM.
The Following 4 Users Say Thank You to JayP For This Useful Post:
I want to see how his stats are affected by things like being forced to play several consecutive games and back-to-backs, and whether or not his season totals are better when he plays more. I have a hunch that his stats are getting worse every year that he plays more than 65 games.
As for other goalies, I don't give a crap unless they are going to play for us.
Yup, no clue what all this before, during and after stuff means or why you have 2 charts
Okay, here's Guy Hebert's relevant years:
Year Team GP
1997-1998 ANA 46
1998-1999 ANA 69
1999-2000 ANA 68
2000-2001 ANA/NYR 54
DURING = stats from 1998-99 and 1999-2000 (years where he played 68+ games)
BEFORE = stats from 1997-98
AFTER = stats from 2000-01
There's two tables to see if there's any impact on fatigue from playing 2 consecutive years of 68+ games and playing 3 or more consecutive years of 68+ games.
I want to see how his stats are affected by things like being forced to play several consecutive games and back-to-backs, and whether or not his season totals are better when he plays more. I have a hunch that his stats are getting worse every year that he plays more than 65 games.
As for other goalies, I don't give a crap unless they are going to play for us.
His stats are getting worse every year he plays more than 65 games. But correlation doesn't mean causality.
And that's a terribly ignorant way of thinking. If other goalies have shown that the year after they get a reduced workload they show no statistical improvement, why should Kiprusoff be any different?
Makes sense to me. Hardly shocking, IMO. Thanks for putting this together.
The real importance of stats can be hard to sort out, so I'm not sure how far to read into this. I don't doubt that your numbers have some cursory meaning - possibly enough to dispel the "myth" - but how far are you willing to take it? Are you saying that Kiprusoff is more likely to have better numbers next year if he plays 70 games as opposed to 60 games? I don't consider that a reasonable conclusion, even based on the above data.
Looking at the statistics more closely, one could go as far as to argue that only "better" goalies will get an opportunity to play 68+ games in a season, so, even more of a statistical improvement in "DURING" and "AFTER" than we see in "BEFORE" might be expected; therefore, the exact opposite of your conclusion could be argued... that "overworking a goaltender" is a real phenomenon, since the negligible improvement in stats from "BEFORE" to "DURING" and "AFTER" are lesser than what we might expect from the "better" group of goalies given a chance to play 68+ a season (especially repeatedly).
It's really all in how you choose to look at the numbers. :|
His stats are getting worse every year he plays more than 65 games. But correlation doesn't mean causality.
And that's a terribly ignorant way of thinking. If other goalies have shown that the year after they get a reduced workload they show no statistical improvement, why should Kiprusoff be any different?
My point is that statistics are only probabilities. You can't say that all goalies do not become over worked based on your numbers.... so, what about Kiprusoff now?
__________________
Ain'T goT no how waTchamacalliT
Last edited by FlamesAddiction; 04-29-2009 at 11:05 PM.
The Following User Says Thank You to FlamesAddiction For This Useful Post:
Makes sense to me. Hardly shocking, IMO. Thanks for putting this together.
The real importance of stats can be hard to sort out, so I'm not sure how far to read into this. I don't doubt that your numbers have some cursory meaning - possibly enough to dispel the "myth" - but how far are you willing to take it? Are you saying that Kiprusoff is more likely to have better numbers next year if he plays 70 games as opposed to 60 games? I don't consider that a reasonable conclusion, even based on the above data.
Looking at the statistics more closely, one could go as far as to argue that only "better" goalies will get an opportunity to play 68+ games in a season, so, even more of a statistical improvement in "DURING" and "AFTER" than we see in "BEFORE" might be expected; therefore, the exact opposite of your conclusion could be argued... that "overworking a goaltender" is a real phenomenon, since the negligible improvement in stats from "BEFORE" to "DURING" and "AFTER" are lesser than what we might expect from the "better" group of goalies given a chance to play 68+ a season (especially repeatedly).
It's really all in how you choose to look at the numbers. :|
I wouldn't say that Kiprusoff played 70 games next season instead of 60 that he'd have better stats. I'd say that he's probably just as likely to put up the same numbers playing either. That's not to say he should play 70+, only that it is more of decision based on how capable our back-up is of playing near Kiprusoff's level.
The small difference in BEFORE and DURING numbers probably can be attributed to the goalies entering their prime. Most of the DURING years were in the goalie's peak years and were probably just starting to break out the year prior, but were stuck in situations where the back-up had to be given sufficient numbers of starts (similar to St. Louis' situation with Chris Mason).
Obviously there's a lot to argue either way, but it's pretty obvious that resting a goalie 5-10 more games isn't going to completely transform his game. At best there will be a moderate improvement in his stats.
Makes sense to me. Hardly shocking, IMO. Thanks for putting this together.
The real importance of stats can be hard to sort out, so I'm not sure how far to read into this. I don't doubt that your numbers have some cursory meaning - possibly enough to dispel the "myth" - but how far are you willing to take it? Are you saying that Kiprusoff is more likely to have better numbers next year if he plays 70 games as opposed to 60 games? I don't consider that a reasonable conclusion, even based on the above data.
It would be more accurate to look at all goalies and compare there stats within different time frames of the same season. Looking at their overall season stats doesn't tell you if maybe they played lights out for the first half and then became average between games 57-72.
Professional athletes are individuals and you can't always use stats to explain things that are unique to them as a human. I remember reading a story about an Olympic Canadian swimming coach who explained why some of his swimmers practiced more than others. He described how each peaked with different amounts of rest. Some performed their best after practicing several times a week, and others with less. As a coach and trainer, he had to find where each swimmers balance was.
My point is that statistics are only probabilities. You can't say that all goalies do not become over worked based on your numbers.... so, what about Kiprusoff now?
Well, yeah they're only probabilities. However, the traditional logic and rhetoric is that over-working a goalie causes them to play worse and playing them less in future years would improve their play. Probability shows that they're more likely to have the same level of play they did when they were playing 68 games or more.
I guess I can't say all goalies do not become over-worked, but it's pretty clear you can't say the opposite as well.
And I don't know what stats you want me to analyze from Kiprusoff. It's pretty obvious his stats are on the decline since he started playing 70+ games, but is that him being over-worked or is it other factors such as worse team defense or Kiprusoff playing worse?
If you have a goaltender that performs at a level that is below 30+ goalies in the league, do you want him starting 75+ games a year for your team?
If Kipper is still top 5 in the league, but it's our defence that's the problem, how bad is our defence getting?
If overworking a goalie truly is a myth, why doesn't Kipper start and play 82 games?
The real question, the one that everyone decides to ignore when discussing goaltending numbers, is the the second one.
What role is the team defence playing in these numbers? Does anyone really think Kipprusoff is the 32nd best goalie in the NHL? Stats only tell a small part of the story.
This team played average defensively on its best nights, and on many nights was among the worst defensive teams in the league. No goaltender was going to put up top 5 stats behind this team.
Overworking a goalie is a myth (with statistical analysis
