[chrisbennet]

There is a fair chance I maybe an idiot but I've rarely worked on a problem where I could:

"thoroughly understand the requirements and truly define the solution approach, leaving the time estimate to be just a matter of counting the generated tasks"

The problems I've been paid to solve are problems that haven't been solved yet or at least I have never solved them before so I would have difficulty making a estimate.

(This isn't meant as a criticism. It's just my experience.)