A Machine To Pick Startup Winners
technologyreview.com
technologyreview.com
That said, and no doubt this will be controversial, I think there's great potential for Moneyball style statistical analysis of programmers. (Though there are significant organizational and cultural issues involved.)
The reason why stats work in baseball, is that there's ample empirical data: baseball players are up to bat, or have a chance to catch a ball again and again. Like baseball players, there are certain actions that programmers do again and again -- namely commits going to production. If one could get accurate stats on a large enough population, with automated metrics for function points and bugs, it should be possible to produce meaningful stats on programmers working with a given technology.
I think the same logic might apply to the OP. Looking at a startup investment through the lens of a credit score makes some sense, but it seems more likely to avoid predictable losses than it is to identify the next facebook or google.
As an aside, there's something I've always wondered about companies who predict the success of other companies...what score did they get from their own algorithm?
You could argue this both ways. The situation at bat varies quite a bit. You can have different pitchers, different lighting, different wind speed and direction, different umpire, different situation on-base, etc... However, this all comes out in the wash because there are so many iterations.
Programmers commit a lot. If you can get reliable data including bug rates for a given programming environment, I think you'd have good data.
Said another way, if Sequoia or Kleiner or Union Square are investing and Correlation then gets to see the deal, I wonder how many deals their algorithm would pass on.
If this was truly algorithmic startup investing to be the lead investor, it'd be much more interesting. But this method of using social proof as the first filter seems like business as usual.
That said, I imagine this is a great approach from a marketing perspective when raising the fund and reaching out to investors (LPs) as it clearly differentiates the fund from a thesis/approach perspective.