I think you really meant to argue that (outside of mathematics) each link in a chain of inference has a small chance of going wrong, so longer chains are exponentially less likely to be correct than shorter ones. This is well known around here, though hubris does take the better of us sometimes!