Yeah, don't lecture me on analytics. I sell analytical software for IBM.
So, you sell analytical software for IBM and you fail to recognize that this type of analysis, even in sports, is here to stay? Interesting.
This paticular analytic seems flawed. By definition- if the worst teams in the league have great Corsi- and a team like NJ who is winning all the time despite **** Corsi- then it is highly overrated as a metric of success, and it is as highly touted on here as WAR is for baseball
That's because based on your posting, you seem to imply that Corsi is a be all, end all metric. It's not, it's a metric created based on the data that is readily available and it's useful, but by no means foolproof (especially when we are only 20 games in). It started simply with shots, now we are seeing other metrics creep in such as quality of competition, location of shots/heat maps, etc.. It's by no means a perfect metric right now, but you are kidding yourself if you think gathering, analyzing and interpreting data like this, and more data as methods to gather it are improved, are a fad.
When Peters was asked about this, and he's as old school as it gets, he said something along the lines of "We use the data to see if there are things we might be missing on our own, and to validate that what we see on our own is backed up by data." (paraphrased).