It’s Sunday night and you’re writing the list for the week.
You’ve learned your lesson from the last hundred Sundays, so this one is modest. Six items. Reasonable items. You’ve padded the hard one with extra time, because you know yourself now. You look at it and feel the small satisfaction of a person who has finally gotten realistic.
By Friday, three of the six are done. One of the remaining three is the hard one, obviously, and it now moves to next week’s list, where it will sit at the top like a landlord.
The standard diagnosis arrives on schedule. You’re undisciplined. You overcommit. You need a better system, a shorter list, a more honest relationship with the word “realistic.”
Except psychologists have been studying this exact gap between the list and the week for almost fifty years, and their conclusion is more complicated than just laziness. The bias has a name, it survives being told about itself, and the fix that actually closed the gap in an experiment is a specific, slightly annoying exercise that almost nobody does on a Sunday night.
The bias has a name and a famous first victim
In 1979, Daniel Kahneman and Amos Tversky gave a label to the tendency to hold a confident belief that your own project will go as planned, even while knowing that most similar projects have run late.
They called it the planning fallacy, and Kahneman’s favorite example of it was himself.
Years earlier, he’d assembled a team in Israel to write a textbook on judgment and decision making. After a year of good progress, he asked everyone to estimate how long the rest would take. The answers clustered around two years. Then he asked the curriculum expert in the room how long comparable teams had taken, and the man realized, apparently for the first time, that a large share of such teams never finished at all, and the ones that did took seven to ten years.
Kahneman recounted the story in a talk published by Edge, noting that this expert had access to all of that history and had still just written down an estimate like everyone else’s. The book took eight years.
A room full of people who studied bias for a living, one of whom was carrying the exact data that predicted the disaster, and the data never touched the estimate. That’s the machine your Sunday list is up against.
Even your worst-case scenario is too optimistic
The experiment that pinned the fallacy down at the scale of ordinary life came from Roger Buehler, Dale Griffin, and Michael Ross in 1994.
They asked 37 psychology students to predict when they’d submit their honors thesis. The average best guess was 33.9 days. The average reality was 55.5 days, and fewer than a third of the students finished by their own predicted date.
An interesting detail worth pondering: the researchers also asked students to predict their finish date “if everything went as poorly as it possibly could.” That doomsday estimate averaged 48.6 days. Reality still beat it. Fewer than half the students finished by their own worst-case scenario.
So this isn’t a matter of shaving your estimates by ten percent. The catastrophe you’re capable of imagining on a Sunday night is still rosier than your actual Friday.
And the pattern scales without limit. The same paper opens with the Sydney Opera House, estimated in 1957 to be done by early 1963 for 7 million dollars; a scaled-down version opened in 1973 at 102 million. In a later study of 258 transportation megaprojects across twenty nations, Bent Flyvbjerg and colleagues found cost overruns in roughly nine out of ten projects. Your to-do list and a national infrastructure budget are failing in the same direction, for the same reason.
More Bolde Stories
Knowing about the bias does not turn it off
Here’s where the planning fallacy separates itself from ordinary overconfidence, and where the “just be more realistic” advice up and dies.
In one of their follow-up experiments, Buehler’s team gave students a computer assignment and, right before asking for a prediction, had one group describe their track record with similar assignments. These students acknowledged, in writing, that they typically finished about one day before a deadline. They were then explicitly told to keep those experiences in mind while predicting.
Their predictions came out just as optimistic as those of people who were never asked about the past at all.
The researchers could see why in the think-aloud studies, where people narrated their thoughts while estimating. About three quarters of those thoughts were future scenarios, little movies of the plan going smoothly. Only 7 percent touched relevant past experience, and a bare 3 percent considered anything that might go wrong.
When people did remember past overruns, they explained them away, attributing them to causes that were external, one-off, and specific to that project. The in-laws visited. The file corrupted. That was then. The past gets acknowledged and then filed under “does not apply,” which is how you can know your history perfectly and repeat it anyway.
One more twist from the same paper: the fallacy is yours alone. When observers predicted how long someone else would take, they leaned on that person’s track record and came out far more conservative. Your friends already know you won’t finish the list. You’re the only one in the room without access to the obvious.
The one manipulation that closed the gap
The reason that study is worth knowing isn’t the failure. It’s the condition that worked.
A third group of students did everything the previous group did, plus two extra steps. First, they wrote down the date they’d finish the assignment if it went the way their assignments typically go. Second, they wrote a short, plausible scenario, built from problems they’d actually hit before, describing how this task would end up finishing at that typical time.
In other words, they weren’t just asked to remember the past. They were forced to connect it, in writing, to the thing in front of them.
That did it. The optimistic bias in that group was eliminated, and 60 percent finished by their predicted time, compared with 29 percent of the control group.
A catch to consider: the exercise made predictions unbiased rather than pinpoint accurate. Some people swung past realism into pessimism, so the average landed true while individual guesses still wobbled. But for a to-do list, unbiased is the point, because the damage comes from the systematic lean in one direction, week after week.
This is the household version of what Kahneman calls taking the outside view, and what forecasting researchers formalized as reference class forecasting: stop asking “how will my plan unfold” and start asking “how do things like this usually go.” Governments now use the industrial-strength version to budget rail lines. You can use the kitchen version on a Tuesday deadline.
So build next week’s list from your record, not your plan
The plan is the trap. The moment you start imagining the steps, you’re back in the movie where nothing goes wrong, and the movie is where the fallacy lives. It’s fantasy land.
So before you write down when something will be done, answer a different question first: when did the last three things like this actually get done? Not when you meant to finish them. When they were finished. Then write one ugly sentence about how this task ends up taking that long too, using problems that have really happened to you, because they are the ones that will happen again.
It will feel pessimistic. The students who did it were more right than everyone else in the building.
And notice what the diagnosis was never about. Not laziness, not discipline, not a character flaw waiting for the right app. If the delay comes with dread, with the task you can’t even start, that’s a different animal with its own research, because procrastination runs on emotion, not scheduling. But the list that shrinks by only half every week is usually just this: a mind that narrates futures beautifully and files its own past under irrelevant.
The list was never unrealistic because you’re broken. It was unrealistic because you wrote it as the person you planned to be, and next Sunday you can write it as the person you have reliably been, who, it turns out, gets a very predictable amount done.
