Quote:
Originally Posted by Buster
Good scientists are often more fascinated by the outliers and the non-correlated results, than they are the garden variety results.
Corsi or whatever flavour is neither right nor wrong nor good nor bad.
It's either useful or it is not useful. I think evidence seems to be mounting that it is limited - not because of the results that agree with it, but because of the valid results that do not.
There is something interesting going on with both the Flames and the Kings vis-a-vis statistics. I'd be more interested in that analysis than a constant defensiveness with regards to the current tools.
What is different about the Flames? Can it be quantified in some reasonable manner? Can that information be used to move on from Corsi to something that is a more complete model?
|
Great post. I have attempted to make this argument a few times, but you have stated it very well here.
The problem isn't statistics, they are what they are. The problem is the analysis and use of said statistics. In my profession, grabbing a stat and drawing a conclusion from it would get you fired so fast you would simply become another statistic. Analysis requires attempting to
disprove the stat, and compare its results with other stats and other forms of analysis, to see if its results can be independently verified. THEN you are in a position to start to draw conclusions.
The 'hockey analysts' learn about a stat and they suddenly think they 'know' something. And to make it worse, many of them get arrogant about it. Drives me figuratively insane.