Quote:
Originally Posted by Inglewood Jack
Goal differential is so close to just outright counting wins that it's about as useful as stating the sunniest days are going to be the ones with the fewest clouds in the sky. The whole point of of the advanced stats is to try and determine which teams are playing the right way to win, whether they're currently winning or not. Those stats clearly have huge flaws right now, but using differential instead is one step away from pure hindsight.
|
It does not matter what the "point" of "advanced stats" is supposed to be. Stats are stats and their correlation with reality is what it is. Just because you call secondary stats "underlying" or "advanced" or "less hindisight" is meaningless.
Corsi is pretty much only ever used for two things: predicting that teams you like are better than their primary stats indicate, or that teams you don't like are worse than their primary stats indicate. This is how you get people trying to argue that somehow the Flames, whose primary stats were indicating a downfall, was actually due to "turn a corner", and this is pretty much what I hear when you suggest that we should be looking at those secondary stats to see if the team is playing well. No we shouldn't. We should be looking at the primary stats, because those "hindsight" stats are always more likely to give the right answer to that question. Of course every now and then the secondary stats will be right and the primary stats wrong, but you should never make decisions with that hope in mind, because the odds will never be in your favor.
Of course fans like us looking for hope in weak stats is harmless. However I strongly suspect that the Flames organization did exactly the same, used the flimsy cover of positive secondary stats to ignore the big blinking warning lights of the primary stats. Of course I can't know this, but I suspect unreasonable faith in secondary stats telling "the real story" about whether or not the team was "playing the right way" is a big reason why the management just stood by while the season wasted away.
If you have enough data you can always create a statistic that weakly correlates with what you want to see. Organizations have been doing this to fool themselves as long as statistics have existed.