I think the problem with taking the moderate-risk-for-moderate-reward strategy, as an early stage investor, is that a large number of your investments are going to fail anyway. And the returns of a company are not linearly correlated with its apparent risk at the early stage: a company might appear 2x as risky, but end up with 200x the returns if it succeeds. Early-stage investors that try to invest further down the risk/reward curve may find that the decreased risk doesn't end up making up for the decreased rewards, and so the fund as a whole becomes unprofitable.

Larry Page was fond of saying that it's actually easier to work on big problems than on little problems, because a.) there's less competition and b.) you can get people to help you. I'm not sure that's actually true anymore - now that everyone wants to be Google - but it's illustrative of the non-linearity of returns.