This post is dedicated to one of the most influential managers I've ever had the pleasure to work with.
I'm of the opinion that the moment you start trying to apply complex mathematical rigour and time to estimation you've entered into a broken system. You're wasting time trying to protect yourself from finger pointing managers or getting a kicking from old school "us or them" lawyer types. None of that adds anything to the project except a big stick to beat people with and make sure they work themselves to death based a a random stab in the dark that is somehow made more realistic by the application if science. Ultimately though if the initial guesswork is wrong, and it's seldom on the nose, the science means sweet FA. So why bother? You're pleasing the wrong people. But then those people hold a lot of power. So as long as we can keep them happy with the minimum of effort life should be sweet.
Ladies and gentlemen I bring you the Estimation 4x4 - guaranteed to knock those super science estimation techniques out of the park.
The Science Bit
Got a project to estimate? I've done your homework for you - 4 people, 4 weeks. BOOM! Mind Blown! Now get on with delivering that project buddy. Oh wait it's a big ass project - 4 people, 4 months. Oh wait you're limited by deadlines and/or resourcing - adjust as needed (3 people, 5 weeks or 5 people, 3 weeks).
Careful now, 9 women dont make a baby in 1 month so don't assume that scaling one side up to cater for the other will get the same results. Theres a balance that needs addressed. This is real science after all - repsect the rule and the rule will respect you.
You Gotta Be Kidding Me
Dont believe me? Try it out next time. Do your pseudo-scientific enterprise-ready finger in the air analysis. But hey see when that project is finished see how close the 4x4 would have brought you to the real result. I bet 9 times out of 10 it was as accurate as your real estimate, except you could have started much sooner....