View Single Post
05-09-2013, 12:34 PM
Registered User
CanaFan's Avatar
Join Date: Feb 2010
Location: BC
Country: Canada
Posts: 3,761
vCash: 500
Originally Posted by RandV View Post
10 people can watch the same game and come up with 10 different interpretations. 'Watching the game' is **** because it depends heavily on the watchers ability to interpret the game, something that's very hard to quantify itself. Especially if it's just a bunch of people on a message board.

Stat's themselves aren't perfect and people can twist them how they will, but at least the same stats are there for everyone to see.

You're not wrong in how you describe watching the game (qualitative data) and using stats -advanced or otherwise - (quant data). However, as someone who works in market research and uses/interprets both regularly, the best approach is to use quant data to interpret and understand what you are seeing/hearing in qual data. Since quant data is often context-poor, it can be mis-used or mis-interpreted quite easily. I find that occurs frequently on this board where advanced stats are used to "disprove" observational data - for example to support that the Sedins played great because their CorsiRelComp score was so high - while watching them play tells you they actually played quite poorly. This is a classic case of cherry picking quant stats and believing that they "trump" observed data, when in fact you can't say one "trumps" the other, you can only use them in tandem to better understand the greater picture.

CanaFan is offline   Reply With Quote