I'm not sure what you mean with the Thomas and Kipper example...But I think in both of there cases it obvious they both couldn't put as high of a save percentage in a big workload years as they could playing few games.
The Thomas example seems rather glaring.
Tim Thomas:
38 GP, .917
66 GP, .905
57 GP, .921
49 GP, .931
It's not at all obvious, you are just seeing what you want to see.
Thomas played 277 more minutes in 2006-07 than he did in 2007-08. That is the equivalent of 4 and a half full games. There is not a chance in the world that those extra minutes had a significant impact on his ability to stop the puck. Let's assume he played those extra 277 minutes in 2007-08, and that he faced shots against at the same rate. For his .921 to drop down to .905, he would have had to let in 42 goals on 143 shots (.706 save percentage). Do you think that would be likely to happen?
Same thing with Nabokov - there is no significant difference between 59 games and 65 games, it's one extra game per month. Hockey Outsider's stats make that pretty clear.
Also, if you look at month-by-month goalie results, they are very similar month-to-month. Goalies don't get worse at the end of the season. In fact, they usually do worst in the first month of the season, before playing themselves into form. That's not what we would expect if fatigue is a very significant variable.
And interestingly Brodeur's highest save% in a full season came in a year he played his fewest games...Maybe just a coincidence? It seems rather significant to me.
Nope. Look at his backup goalie, Mike Dunham. With Dunham behind him, Brodeur had two of his lowest seasonal games played totals. With much worse goalies behind him, he played a lot more games. The talent of the backup goalie is the most significant variable in determining the number of games played by a starting goalie.
Take, for example, Thomas in 2005-06. He played more games because his backup was terrible (Hannu Toivonen and his .879 save percentage). Thomas played fewer games in 2006-07 despite playing much better, and is on pace for even fewer games this season despite playing like one of the top goalies in the league, simply because Alex Auld and Manny Fernandez are much better backups.
Nabokov is another great example. He played with Toskala in 2006-07, and they split starts. Then he played with Dmitri Patzold and Thomas Greiss, and he played 77 games and was nominated for the Vezina. His stats are almost identical both years, so somehow having worse backups means you are a better goalie?
Like Seventieslord says, if fatigue effects exist, and goalies play worse the more games they play, then there is some threshold where the team is better off having the backup goalie in the net. The better the backup, the more times he is likely to exceed the threshold, and the result is fewer games for the starting goalie.
But I think you would have to look at the numbers on a goalie by goalie basis....If Brodeur plays 70+ games and then he plays 50 games is there a difference in HIS save%?
That is the real question.
Exactly. And this year Brodeur has only played 22 games on a much better team, so naturally he has a much higher save percentage than he did the last two years:
2006-07: 78 GP, .922
2007-08: 77 GP, .920
2008-09: 22 GP, .922
Hang on a second....