Assuming that those numbers are known precisely, the 5%/$200M is clearly the better option. But in practice it's basically impossible to distinguish a 1% succes

Assuming that those numbers are known precisely, the 5%/$200M is clearly the better option. But in practice it's basically impossible to distinguish a 1% success rate business from a 5% success rate business. On the other hand, it's pretty doable to estimate best-case outcomes based on the size of the market being addressed and some assumptions about margins and market share derived from the business model.