Regression Analysis. I used to run regressions all day long when I worked at Booz-Allen — trying to unearth manufacturing cost behaviour as a function of scale, complexity, etc. Back in those days I used a calculator or a package that was called StatGraphics on the pc, which i see is still around. I remember forcing Booz-Allen to buy IBM PC ATs so that we could run StatGraphics faster.
I just did my first regression in years. Just used the data analysis pak that comes with Excel. Really easy. Flipping variables and observations in and out of the set is not very easy, but the basics work fine. Charts aren’t very exciting either. I bet there is a better add-in to use — Google lists quite a few.
I was analyzing VC returns for most recent funds. What I learned — fully half the variation in returns to date is unexplained by quantitative variables (at least the ones I had at my disposal). Sector focus? Quality of VC staff? Hard to know. Of the rest, returns are most correlated with number of deals, and % of fund invested. Still wrapping my head around this but seems to say: don’t by shy about putting your money to work, and spread it around across a lot of deals (portfolio theory at work)?