Moneyball Update | Hacker News Reader

Moneyball Update

sports.espn.go.com

8 pointsfigital5 comments

@lionhearted where do you get your raw basedball data? it's certainly a fun way to learn/teach data mining but the folks at the top certainly are not going to be publishing their formulas. ;)

I like the discussion at sonsofsamhorn.net - a Red Sox forum, but a hell of a lot of smart people there. They get their data from the regular sources, there's lots of citing in threads. I don't datamine it out myself - nowhere near enough free time.

Thanks for the rec. I just noticed Sean Layman's Baseball Archive which looks to have individual plate appearances from 1973-2008 and is available in a SQL-ready format:

http://baseball1.com/content/view/57/82/

This looks like it might be the same DB:

http://sn.im/baseball-sql

I love baseball - because of its unique start and stop nature of relatively isolated events, it's one of the easiest sports to do lots of interesting analysis in.

As for Moneyball - first, high onbase is now getting much more recognition and payment, and he's a bit of a victim of his own fame. His type of star he could get undermarket, he now has a lot more competition for.

Second, the A's were thought to be able to contend this year - they even traded for a top player (Matt Holiday) to try to put them over the top. They'll probably contend next year in a weak Western division.

Most interesting - defense is probably the new onbase percentage. It's pretty easy to track offense in baseball, and separate good and bad offense. Also easy to track pitching. Defense? Not so much. There's not as many numbers kept, and all the major defensive systems have serious flaws. For years, good offense/bad defense players did really well financially and cost a lot. It's changing a bit now, and better GM's are looking for average offense/great defense as the new kind of bargain.

I was confused for about 3 minutes because I kept reading "Monkeyball".

Reply on news.ycombinator.com