[bsder]

> If you could increase the odds of winning a game by 5% using a million bucks of IBM 7094 mainframes and a million LoC in COBOL then gramps would have done it and made a profit to boot and maybe you can make a couple dimes by modernizing the hardware, but I assure you the algorithms and constraints are baked into the business logic cake and the win/loss percentage won't change.

"Moneyball" demonstrates that your assumption is incorrect.

In addition, you have to collect data, first, before you can analyze it. The limitation blocking "Moneyball" from happening was the fact that the statistics didn't exist for a long time.

Several of my friends were among the people who originally helped Bill James by scoring all the games they went to and then sending the papers in the mail.

Sometimes, something really does require "one clever person" in order to put it all together.