The article doesn't contain much information on what this "moneyball" actually is.

[Also, the author might like to know that saving powerpoint files as png files (i.e. "save as" then select .png from pull-down), rather than taking screenshots gives a more readable result.]

The core moneyball idea, for me at least, was identifying an undervalued potentially winning strategy with the help of data and then using data to identify players to best cary out that strategy. Interestingly, in baseball, the A's are back to a homerun hitting strategy because moneyball caused their original strategy of getting on base to become overvalued, leaving other market opportunities. So avoiding homeruns wasn't really the core idea. And it should be noted the As have never won with this strategy. They won with steroid taking supermen.

"Data driven selection model" sounds nice and all, but it is hard to quantify something as abstract as an early stage start-up. Many companies are barely more than a couple of guys and an idea...

Compared to Warren Buffett that mostly invests on stocks, something that can easily be broken down to mere numbers.

Every little bit helps, especially when talking about something as abstract. Besides, most investors do try to get as much data on the team background, market, etc in their decision process.

Of course, i don't expect software algo's to choose startups for investment (like they mostly do with stocks) any time soon.