Exactly my point. Come up with some hockey stats that predict anything at all and I'll definitely listen. Currently, the vast majority of stats talked about in hockey are more deceptive than they are helpful. No stats at all are better than bad statistics. Just about every stats book in the world that's been written since the 80's devotes entire chapters to that point.
Statistical analysis or what have you of sports is naturally going to be uncertain because there are loads of uncertainties. It has to be looked at in that context, and the question that has to be asked is: what's going to be predictive? If it's predictive, how predictive is it? And the truth is, there's not very much that's predictive in hockey. A little, but not a lot. The games still have to be played. Regardless, there are plenty of errors that people make when looking at 'their' statistics versus the newer stats.
Hockey is different from even basketball, in terms of statistical analysis. There's so little that is discreet that it'll take some extraordinary evidence to support just about any statistical claim. The failure of most amateur hockey statisticians to even try to come up with even average evidence for their claims make hockey statistics a running joke.