One issue I have with these sorts of studies is the short-term nature of the tasks the programmers are given. It's easy to complete short assignments quickly i

One issue I have with these sorts of studies is the short-term nature of the tasks the programmers are given. It's easy to complete short assignments quickly if you know you'll never need to maintain the code. A lot of the time the programmers that I would class to be in the 10x range don't complete the early tasks in a greenfield project much faster that an average programmer - however the choices they make, the way they write their tests and the architecture they use results in a codebase where their next tasks are as easy for them as the previous ones. Come back to the project two years later and they're still finding new tasks easy to complete.