Dr. Dunk --- thanks for the tip! I briefly perused the site you posted. Its stimulating (or should I say simulating?)and daunting as well. I do like the idea of exploring the ideas/observations/claims made by folks on the board using empirical methods, when applicable (what can I say - I'm a social scientist). But as you say, time (the lack thereof) precludes me from actually pursuing this all too much. maybe when I get tenure (haha). Good topic and nice insights. I like this a lot more than over-analyzing/over-hyping trade "possibilities"
Its a bit like which came first the chicken or the egg argument. Did Anderson get 20 minutes cause his game was AWOL or Did Andersons game go AWOl cause he only got 20 mins. I think also the 20 mins was a small reflection on Mobleys good D on Jones
Outseam, I'm in the process of doing just that. I've just got a ton of crap going on though. I may start a thread in the BBS Hangout part of the BBS to discuss this. Or I may put up a temporary BBS on my own server. We'll see... I want to start with an efficiency stat. How strong is your statistical knowledge? ------------------ trade them all and fire the coach.
That would be an entertaining side-project. My stats knowledge is fair to middlin' --- I'm strong in practical applications but a little bit weaker on the theoretical underpinnings of the whole venture. Probably the most useful techniques for a project like this would be multiple regression, maybe path analyses and/or structural equations modeling - all of which I'm competent at. Now, if anybody knew dynamical systems techniques... Don't know much about baseball, but it does seem that part of that tradition involves a fascination with keeping stats. IMO, this is not the case with basketball. Why might this be? Is there a sense (and perhaps a reality) that the game is too fluid and dynamic to be captured in any convincing sense by statistics? I like the idea of an efficiency stat. Perhaps we should carry on our conversation of board so's we don't bore folks too much. Another thing I've been thinking about are "best line-ups" (e.g., most cohesive offensive line-up). I wouldn't want to go the fantasy basketball route and use some formula that essentially adds up players' contributions. The ideal would be actual team productivity during the span of time that say Francis, Mobley, Shandon, et al. played together. Of course, that would require data collection beyond that of the average, and even rabid, fan. I wonder if the Rockets themselves collect that kind of detailed analysis...? Outseam
http://mathworld.wolfram.com/DynamicalSystem.html [This message has been edited by BobFinn* (edited February 24, 2000).]
