A Final Thought On Software Budget Estimation

Consider this, several years back I was having a problem with a building automation system. I spent several days in the field trying this or that to solve the problem. I finally stumbled on the solution, very simple one at that; modify one line of code and just like that the problem was solved. Upon returning to the office my boss asked if I had fixed the problem, to which I replied, "Yes, I just had to change one line of code". This of course we met with a very disgruntled response, "But you've been gone for two days!" Looking back at situations like this really begs the question, how do you apply an accurate estimating system to this type of problem? A two minute fix, with 15 hours and 58 minutes of contingency? Good luck using that estimating technique and actually winning a competitive bid! Bottom line, you can't win them all, just hope you win more than you loose.