Posted by jetwildcat on 7/2/2012 11:14:00 PM (view original):
billyg, i have to ask...how much analyzing time do you end up saving because of your comp sci background?
well, if what you are asking is, what does my PROGRAMMING save me -- nothing. i purposefully do NOT farm the data of seasons of games, and reverse compile the secrets of the game. its very difficult for people to do - many have tried, and failed. they simply run simple regressions and go wow, shot blocking is REALLY important. well... no, no its not. the only way those simple regressions work is if *recruit generation ratings are independent variables*, and clearly, they are not (ath/def, sb/reb, both have extremely tight correlations, to name a few).
however, if you pair statistical analysis, with a strong understanding of the game, you could probably get exact values on things you shouldn't be getting exact values on. and anyway, i feel like i already know what ratings are valuable for, how valuable they are, and how they complement each other.
on the other hand, i did go to CMU, which is about as good a theoretical computer science program as you will find in the world. the problem solving and analytical skills they work on there are really valuable for tons of abstract stuff. i can't tell you how i could study a team and try to predict the outcome of a game, then study the outcome, and draw valid conclusions - especially with the incredible number of variables involved. but the problem solving and analysis i built there definitely helped! probably helps too, that i didnt know a damn thing about basketball :) i didn't have paper and pencil tricks or programming tricks and i didn't keep track of anything. im sure programming would have sped up the process - but i don't think it would have been as much fun. i do think most people would benefit from that kind of analysis, but i think it would take them longer, and most people probably couldn't get to the same level of detail. so in that regard, the comp sci definitely helped.
i would *REALLY* like to write a sim AI for this game though, but at that point, i may as well just write my own sim engine and be done with it...
7/3/2012 12:27 AM (edited)