You'd be better off noticing the patterns in his misestimation and adjusting his estimates accordingly in your own head. Better yet, encourage him to become a

You'd be better off noticing the patterns in his misestimation and adjusting his estimates accordingly in your own head. Better yet, encourage him to become a better estimator. The book Software Estimation: Demystifying a Black Art presents several strategies for developing this skill. For example: before starting a project, classify it as “small”, “medium”, “large”, “x-large”, etc. (If you don’t understand what you’re supposed to be doing well enough to even make an estimate, you have a bigger problem.) After the project is over, record how long it took. After a few months of work you should be able to see what ranges of time are associated with what intuitive size estimates, and then you can translate your intuition into numbers.