The hardest single part of building a software system is deciding precisely what to build.
I don't try to predict the market. I just try to buy stocks that are cheap.