posted at 3/2/2011 8:29 AM EST
Back when I was a statistical philistine, I loathed baseball. It took going to a live game in Toronto this past summer for me to truly appreciate the cerebral nature of baseball. For a time, it became my obsession, until my A.D.D. carried my attention elsewhere (Ooo, shiny!). Mostly, my discovery of sabermetrics led me to ponder how this arcane science could be applied to my first love, hockey.
I came across an interesting site I thought I'd share, especially if there are any fellow forumers out there with a love of stats, statistical analysis and sabermetrics. Hockey Analytics: www.hockeyanalytics.com/
" target="_blank" title="Hockey Analytics"> www.hockeyanalytics.com/
Perhaps, as this author says, "it isn’t easy to use statistics to gain an insight into the game of hockey. The first problem is the game does not lend itself to a meaningful statistical summary. A second problem is that nobody has ever spent much time thinking about what statistics should be captured and how. Finally, the game is “chaotic” and full of interrelationships."
Baseball seems to lend itself more easily to deep statistical analysis, as there are more "individual" actions - at bats, for example. Hockey is a bit more slippery, and even "individual" actions in hockey are comprised of several external contributing factors and components. In that respect, baseball is more "mathematizable" than hockey - but it did take 20 years for baseball sabermetrics to reach adulthood. Hockey sabermetrics are certainly in their infancy, and maybe eventually a systems is constructed that takes peripheral stats like takeaways, giveaways and scoring chances, and plugs them into the metric in a relevant way.
Part of the problem is that key metrics have only (relatively) recently started to be recorded (like time on ice), while others have changed over time (there used to be one assist per goal awarded...now there are two), which prevents anyone from delving too far into the statistical pool of the past to seek out trends in the numbers. The author of hockeyanalytics (Alan Ryder) has done some good work so far, and he has some external links to other analysts and win probability models, if anyone is interested in reading further, or getting a "second opinion", as it were.
Perhaps "Moneypuck" remains forever elusive to the scrying statistical eyes of the sabermetrician(s). Interesting stuff for anyone who's a numbers/stat geek.
Cheers, and enjoy