I did not achieve a certain milestone at work that I had planned to reach by the end of the week. This means I have failed at estimating how much time I needed to reach it. This is bad because it happens fairly often, and the converse (finishing sooner than expected) happens rarely. Despite everything I know about fallacies and biases with predicting schedules, I keep predicting less work than is really required.

My work is programming. It is notoriously hard to quantify (or the managers could make predictions themselves without asking me). If I can't modify the intuitions I feel, I may have to try just multiplying my feelings by a factor - e.g. adding 100% time to each estimate.