Estimate frequently, quickly, and move on to execution. This will allow you real-world information to make a better estimate next time. Lots of little fuzzy gue

Estimate frequently, quickly, and move on to execution. This will allow you real-world information to make a better estimate next time. Lots of little fuzzy guesses which trend in the right direction. I believe the complaint about "not getting any closer to resolution" is a red herring. The value in the exercise is creating a shared mental model, not creating COCMO VII. There is no perfect estimation model, mainly because your team is not full of robots. We screw up when we get up from the IDE and start thinking of technology development in the same way as we would using the Jquery framework. People do not work like machines.