Planning fallacy
The planning fallacy is the tendency to estimate your own work from the best case, and to keep doing it after every previous estimate came in short.
Kahneman and Tversky named the planning fallacy in 1979: people predict how long a task will take by picturing it going well, and stay optimistic even when they know similar tasks have overrun. The striking part of the finding is not the optimism itself. It is that personal experience of overrunning does not correct it, because each new plan gets judged on its own details rather than against the record of the last ten.
The textbook demonstration is a student estimating when a thesis will be done. The textbook large example is a public building: the Sydney Opera House was projected at seven million dollars and four years, and finished at one hundred and two million and ten. The same pattern holds at every scale in between, which is why it survives as a rule rather than as an anecdote.
Kahneman's own correction is reference class forecasting, which means predicting from how similar past work actually went instead of from a mental walkthrough of this one. The version most people can run is cruder and works well enough: take the estimate, note how wrong the last few were and in which direction, apply that ratio, and leave slack for the things that were never on the list at all.
Why it matters
An overrunning day does not fail evenly. It fails at the end, and what sits at the end is whatever had no external deadline, which is usually the people. A plan built out of best cases spends its optimism first and pays for it out of the relationships.