View Single Post
02-04-2012, 10:05 AM
Registered User
Join Date: Jan 2006
Posts: 4,254
vCash: 500
Attempts at quantitative analysis in Hockey generally suffer from two problems. First it is much more difficult to decouple the effect of one player from the other players on the ice. Second, the statistics that have been traditionally recorded (Goals, Assists, etc) cannot be used to generate statistics that are as predictive as those that have been found for Baseball.

The second problem is pretty easy to solve, hire people to watch every game and record different statistics that are more meaningful.

The first problem is a bit more difficult but many scientific disciplines have over come much much more complicated (and complex) problems. With a background physics/biology/economics I can think of at least a half dozen techniques off the top of my head that have the potential to work.

Despite their popularity today, Sabremetrics still struggle to say much about pitching, one of the main elements of the baseball. Furthermore, 15+ years ago there was a lot of debate in the Sabremetrics community as to whether or not a useful statistic could ever be arrived at for fielding. People tried everything with the traditionally collected statistics and got nothing (range is garbage IMO). But then some people started to collect new positioning data and they basically broke the problem open.

IMO, high level quantitative analysis may or may not work in Hockey. There are a lot of challenges and there are a lot of techniques that may or may not be able to meet those challenges. What I can say with absolute certainty is that any person that tells you that they definitely can/cannot work is either full of **** or the most intelligent human being on earth.

almostawake is offline   Reply With Quote