Here's first round save percentage back to 1983-84, normalized to a 900 save% league:
SV% | 907.7 |
STDEV | 29.3 |
Now if I take a 90.77% weighted coin and run 10,000 trials for each of the 400 goalie-series we had from 83-84 to 08-09, I get:
SV% | 907.7 |
STDEV | 23.3 |
That says luck is 63% of first round save percentage, and non-luck factors are 37%. So if goalies all had identical talent, we'd find that 23% of them exceeded Montreal's 925 save percentage in the opening round; in reality, 28% of goaltenders exceeded 925 in the 400 series since the NHL started recording shots against.
To put that in annual terms, we would have expected 3.6 of 16 first round goaltenders to hit this standard; in reality, 4.5 goaltenders did. So talent gives us not even one additional high-performing goaltender per year.
Vic brings up an important point below – the league is very different after 1990:
STDEV | LUCK | Non-Luck | |
1999-2009 | 25.4 | zero | |
1990-2009 | 26.1 | 22.7 | 13.0 |
1983-2009 | 29.3 | 23.3 | 17.7 |
Extreme playoff goaltending performances are much rarer over the last 20 years than they used to be. Ken Dryden really was that much better than everyone else, but no such goaltender exists today. Since 1990, we expected to see 3.6 save percentages above .925 each year from pure randomness; we got 3.9.