It depends on your view of what makes the world a better place.

For tech startups, the distribution of world-changing impact probably closely follows the distribution of investment returns. If it's extremely bimodal, then it makes sense to concentrate on generating big wins.