You're previewing the Bennett demo family. Numbers, schedule, and kids are sample data.Start your own
Loading page…
Researchers asked students how long their big project would take, then asked for a worst-case guess too. The real answer came in later than the worst case. There is a name for this, and once you know it you can plan around it.
You look at the assignment. You think: that's an hour, maybe. Three hours later you are still at the table, faintly annoyed at a version of yourself who lived earlier that afternoon.
The interesting part is not that you were off. It is that you are off in the same direction every single time, and that knowing about it does not automatically fix it. That is unusual enough to have been studied properly, and it has a name: the planning fallacy.
In 1994, three psychologists — Roger Buehler, Dale Griffin and Michael Ross — asked 37 university students to predict when they would finish their final-year thesis: the biggest piece of writing any of them had ever done, and one they had months of experience of already.
Then the researchers asked two more questions, and those two are what make the study famous.
Then they waited, and wrote down what actually happened.
Each bar is an average across those 37 students. Read them left to right: their sunniest guess, their real prediction, their gloomiest guess — and then the last bar.
| Estimate against reality | Days |
|---|---|
| Best case | 27.4 |
| Predicted | 33.9 |
| Worst case | 48.6 |
| Actual | 55.5 |
Sit with bar 3 and bar 4 for a second.
Bar 3 is what these students said would happen if everything went badly. Bar 4 is what really happened. Reality landed past the disaster scenario — and only about 30% of them finished inside their own prediction.
They were not being silly. They knew the work, they had done years of assignments, and they were the world's leading experts on their own habits. They were still 21 days out.
The trouble is not that people fail to imagine things going badly. They imagine it, write down a number for it — and reality still comes in later.
Look at the four bars again. What is the most surprising thing in that chart?
When you estimate a job, you do not consult your history. You build a little film of how it is going to go: sit down, find the notes, write the first paragraph, keep going, finished.
That film is honest and specific, and it is also the version where nothing goes sideways. There is no scene in it where the printer is out of ink, or your sister needs the table, or paragraph two turns out to need a fact you do not have. Those things are not in the film because they are unpredictable — you genuinely could not have scripted them.
But something interruption-shaped happens every time. Each individual one is a surprise; the fact that there will be some is not.
And here is the move that keeps the whole thing running. You have the evidence. You remember that the last project ran long. You simply file it away as a one-off — that week was unusual, my computer was playing up, that won't happen again — and predict from the film instead.
Every past overrun had a specific reason. So each one gets excused, and the pattern they add up to never gets counted.
In the same set of studies, the researchers tried something else. They gave people the details of someone else's task and asked how long that person would need.
Those outsiders did not underestimate. If anything they guessed high — and the reason they did better is the interesting bit: they leaned on past experience, which the person doing the task had and did not use. From outside, you cannot see anybody's film. All you have is the boring outside record of how these things usually go, so that is what you use.
Which means the skill is not missing. It is switched off, specifically and only when the task is yours.
Your brother says his science project will take two evenings. You reckon four, and you turn out to be closer. Why were you better at guessing his time than he was?
Six things people try. Three of them are supported by the research you have just read; three feel productive and do nothing to the number. Sort each one and read why.
Tap an item, then tap where it belongs
You do not need a research team. You need three lines in a notebook.
After about five entries you will have something no amount of thinking can give you: your own real record. If your reading assignments have run 40 minutes over three times, your next estimate starts from 40, not from the film.
And notice what this is not. It is not working faster, and it is not being harder on yourself. It is telling the truth earlier, which is a much easier thing to do.
You have a book report due Friday. Your notebook shows your last three reports took 3, 4 and 4 hours. Sitting down on Monday, you feel like this one is a 2-hour job. What should you plan around?
You’re previewing as a parent — nothing here is recorded.
Buehler, R., Griffin, D., & Ross, M. (1994). 'Exploring the planning fallacy: Why people underestimate their task completion times.' Journal of Personality and Social Psychology, 67(3), 366-381. Study 1: 37 psychology students predicted an average of 33.9 days to complete their honours thesis and took an average of 55.5 days; their optimistic-scenario estimate averaged 27.4 days and their pessimistic-scenario estimate 48.6 days; about 30% finished within their own predicted time. Same paper, Study 5: observers predicting another person's completion time did not show the optimistic bias and made greater use of relevant past experience than the people doing the task did. The term 'planning fallacy' was introduced by Daniel Kahneman and Amos Tversky in 'Intuitive prediction: Biases and corrective procedures' (1979). Kruger, J., & Evans, M. (2004). 'If you don't want to be late, enumerate: Unpacking reduces the planning fallacy.' Journal of Experimental Social Psychology, 40, 586-598. Participants prompted to unpack a task into subcomponents gave longer estimates, and in two of the five experiments measurably less biased ones. The book-report example and the notebook figures are illustrations, not study data.