> what's the chance that the average person is going to actually have any of these complications?
Yes, that's the question we'd all like to have the answer to.
But we don't know how many mild or asymptomatic cases have resolved, so the deaths/(deaths+resolved) calculation is a known divided by an unknown.