Research library
Method
Estimation Calibration
Kahneman, 1979
Most people underestimate task duration by a stable factor.
The planning fallacy: predictions about how long a task will take systematically ignore base rates from similar past tasks. The fix is reference class forecasting using your own history.
What the research shows
Buehler, Griffin, Ross (1994) found students completed thesis work an average of 22 days later than their best guess.
Where evidence is weaker
Calibration only works when task types are tagged consistently. Mixing 'write report' and 'fix bug' under one label hides the signal.
How Focus Capsule applies it
Each MIT carries an estimated duration. Over time we report your personal mean ratio of actual to estimated, broken down by task type.