Quote:
Originally Posted by Jay Random
It's a stat derived from goals allowed vs. quality of scoring chances. Both of those numbers, in this case, are based on a much larger sample size than nine.
Whereas this number is, in fact, based on a sample size of nine. (And it reduces every game to a single ternary value, good/bad/indifferent, which means that all the relevant data within each game is thrown away.)
Funny how you use the ‘sample size’ argument when it suits you, and totally ignore it when it does not.
|
Since you've somehoe turned this into an ad hominem - let me put it like this:
GSAA over a large sample size is probably useful to prescribe future results, although goalies still have high variance. It is probably the best stat to compare goalies' value.
GSAA over a small sample size sort of describes past results, but can be easily skewed.
QS over a multi-season sample size is useful to describe (but not predict) consistency. It won't tell you if a goalie can pitch 20 shutouts in a season, but it can tell you if a goalie can perform to a league average level with better consistency than other goalies.
QS over a small sample size is ONLY to describe the results over a small sample size, while taking out the "did the skaters score more goals than the goalie allowed" part out of wins - nothing more, nothing less.
They all describe different things. Since we're nine games into Markstrom's season, I think QS is the only stat useful to describe those nine games.
I am not using not to predict the next 50 games.
I am simply saying that Markstrom has not been very consistent so far, across nine games, that he deserves to be placed on a pedestal.