"New opinions are always suspected, and usually opposed, without any other reason but because they are not already common."

"The final test of any statistic is whether or not it correlates with winning."

"Any statistic the meaning of which can be expressed in understandable terms in a common English sentence is always to be preferred, other things being equal, to one which cannot."

"Good sabermetrics respects the validity of all types of evidence, including that which is beyond the scope of statistical validation."

"Opinions are great—don't get me wrong. They're great for starting research projects. Then you go study and see if you can prove the opinion or not. But when placing multi—million dollar bets on future outcomes, opinions are wholly unsatisfactory. Opinions as conversation starters are fine. Opinions as conclusions are very bad. I started research projects to discern the objective “why.” I wanted to know why certain teams won and why other teams lost; why certain drafts produced big stars and others didn't. This was the naïve question at work."
*****
That's a pretty good introduction to my thoughts on the hockey metrics issue, which is being talked about here, here, here, and here. I suppose I would be considered a "stats guy," but I don't create any of my own and I have my own concerns about how they are being introduced to hockey. About two years ago, I started working on a post on the issue, but put it aside because it was consuming me. It was huge, and I just couldn't narrow down or focus my thoughts. Thankfully, the posts by David, Lowetide, Dave, and Matt have covered a great deal of territory, and I can now toss in my own two cents, mostly centred around what I would call "first principles." So here goes.


Hockey Metrics

State of "traditional" hockey statistics: in dire need of a Cartesian exercise in methodological skepticism.

Goal of "advanced" hockey metrics: the elimination of orthodox, unsound opinions in favor of objective, measurable truths about the game of hockey.

Basic needs to that end:
1) Need to identify what the game is about (runs scored and runs prevented)

2) Need to identify its particular qualities.

3) Need to identify which strategies are the most conducive to success.

4) Need to identify what is common between the “best” players and "best" teams.

5) Need to identify which current statistics are useless, based on above criteria.

6) Need to identify which current statistics are useful, based on above criteria.

Needs to push advanced hockey metrics to next level:
1) Need a Glossary of Terms defining advanced statistics in understandable and accessible language.

2) Need to identify standardized benchmarks of quality (and non-quality) for advanced statistics (like .300/.400/.500 in baseball).

3) Need more carrot, less stick in explaining advanced statistics to doubters, non-believers or newbies.

4) Need to make advanced statistics more publicly available for all to access and play with.


Not the compromised second draft:
• Matt Fenwick
• Matt Fenwick
• Tyler Dellow
• Vic Ferrari
• Gabe Desjardins
• Gabe Desjardins
• Alan Ryder
• Chris Boersma
• Chris Boersma
• Dirk Hoag


There you go. Basic stuff, really, and much is already being done (Matt's post from earlier today being a fine example). But at least it's down and done. Plus, now I can post the final proof that I took too many philosophy classes in university: an image I made some time ago linking Plato and Descartes to Bill James. I'm a DOOORRR ORRRRKKK.