"Adding up estimates rarely work when you end up with more than a few tasks. Instead, figure out which tasks have the highest uncertainty – those tasks are basically going to dominate the mean time to completion." -- Erik Bernhardsson explains why you should worry about risks and tackle them first, as they can have a drastic impact on your time estimations. A good lesson to take from it: teach software engineers how to do proper risk management, delay work around areas of comfort, and create risk-based milestones with adjustments to the time estimations.