Quote:
Originally Posted by V
Do you think that's any different? Practically every scientific experiment I've ever seen is using current data to predict future results.
Are you 8?
|
9 actually, can't wait to turn the big ONE-OH!
The point is this: advanced stats (the way they're used in regards to the NHL) are essentially an unproven hypothesis. People trot them out like they are of clear value and credence, but they simply represent an unproven notion of "what-if". They give no answer.
If ok with someone saying "Hey, we should sign _____, he's this this and this, and his advanced stats look good." There is nothing wrong with using them in that way.
The problem lies when you do what the two bloggers did in the links that 19Yeezie19 provided, and say "Contrary to any visual data or otherwise, this player is actually good/bad at this, because: advanced stats." They simply do not work that way. People who use advanced stats want to use them that way, but it's completely ignorant of their purpose or value.
You don't have to look any further than the lists provided, as has been pointed out, where borderline AHLer's have better "advanced stats" than some of the best all around players in the game. If you're going to say "Jack Johnson is bad because: advanced stats" then you need to be prepared to be totally ok with saying that Jordan Eberle is better than Steven Stamkos, Henrik Zetterberg, or Jonathan Toews. Now, those numbers were taken from the end of last season, but really, has he ever been better than any of those three? Ever? In fact, according to those numbers, the Oilers #1 line was the best #1 in the ENTIRE NHL. People can praise Corsi all you want, but when you say a player is bad because of it, you also should look at which players are GOOD because of it, and how good it says they are for that matter.