[FoRK] Baseball was Re: Mining humans was Re: Big Data
Gregory Alan Bolcer
greg at bolcer.org
Mon Feb 6 14:19:54 PST 2012
I like baseball too, but when I watched the movie, I didn't think of it
as a baseball movie (as much as the story just being set in baseball).
The most interesting parts I liked were a) "google analytics" for
baseball references and b) the organizational pushback on trying a new,
data-centric strategy and a whole game theoretic type approach.
The old equation is something like this:
Finite set of cash / 25 member team roster with 9 player/position pigeon
holes to place them in. Each team spends as much as they can in each
position to get the best player they can--drawing from the same pool.
For each team, spending more on one position means spending less on
others and teams can only afford so many good players. The games where
they put the most money into the team roster, have better players, thus
win more (and sell more tickets and hotdogs etc).
The new equation is, that you optimize key statistics across the team to
optimize the statistics you need to win the most games across the whole
season. It's just an offshoot of global optimization where money isn't
the key metric. They field their team based on other statistics that
are more tightly correlated with scoring runs. The teams that have the
best complementary statistics are the ones that score more and thus win
On 2/6/2012 1:04 PM, Marty Halvorson wrote:
> I love baseball. When I was in middle school, I lived 1 block from the
> Minneapolis Miller home field (around the early to mid 50's). Spent many
> a fine afternoon hanging around outside the field waiting for a foul
> over the fence. When ever I got one, I could turn it in at the ticket
> office and get in free for the rest of the game.
> "Moneyball" is probably my favorite for 2011 also. I thought the movie
> was more about how the equation was changed from spending for the big
> names to spending for players who could score runs (as evidenced by
> their on-base percentages), especially the undervalued players. The
> unfair advantage being that some teams had owners willing to spend big
> bucks on big name players.
> But, maybe Greg, that's what you were saying.
> Marty Halvorson
> FoRK mailing list
greg at bolcer.org, http://bolcer.org, c: +1.714.928.5476
More information about the FoRK