How to guess how long a task will take, and why we guess low
We guess low because we picture the plan going right and set aside how similar tasks went before. In a classic study, students expected to finish their theses in about 34 days and took about 56. To guess better, start from how long the last similar task really took, time a task once, break big ones into parts, ask someone else, and compare your guesses with what happened.
Coming soon to iPhone and iPad
On this page
Why we guess low
In 1994, the psychologists Roger Buehler, Dale Griffin and Michael Ross asked final-year students when they would hand in their theses. On average, the students’ best guess was 33.9 days. They took 55.5. Even their “if everything goes as poorly as it possibly could” estimate, 48.6 days, was too short, and fewer than a third finished within their own best guess.
The researchers then asked people to think aloud while estimating. Most of what came out was about the future: how the work would go, step by step. Very little was about how similar tasks had gone before. And when people did recall past tasks that ran late, they tended to explain them away as outside, one-off causes: the kind of bad luck that surely would not happen again. This time, they felt, would be different.
Two more pieces make the bias stubborn. Research by Roy and colleagues suggests that our memory of how long past tasks took is itself too short, so even honest recall starts from a low number. And the bias is about our own tasks: in Buehler’s fifth study, observers predicting other people’s completion times overestimated them.
Five ways to guess better
- Start from the last similar task, not from the plan. Ask “how long did this take last time, from start to finished?” before you ask “how long should it take?” In Buehler’s fourth study, people who were told to connect their prediction to relevant past experiences made predictions without the optimistic bias.
- Time it once. Since memory runs short, measure. Note the time when you start and when you finish, or run a timer. In a 2018 study by Kevin Thomas and Cornelius König, people’s predictions for a task were less biased when they had done the same task before.
- Break big tasks into parts. People tend to allot less time to a whole task than to the sum of its pieces, which researchers call the segmentation effect. Listing the parts brings the hidden steps into view. It takes effort, so keep it for tasks longer than an hour or two.
- Ask someone else. A colleague or partner does not share your rosy plan and tends to think of what went wrong before. Their guess is a useful check on yours.
- Count the interruptions in. Your estimate is for uninterrupted work, and real days are not. If you are rarely left alone for more than half an hour, a two-hour task does not fit in a two-hour gap.
Keep a one-week estimate log
The fastest way to learn your own bias is to write it down. For a week, before each larger task, write your guess; afterwards, the real time. An example:
| Task | Guess | Actual | Ratio |
|---|---|---|---|
| Weekly report | 45 min | 70 min | 1.6 |
| Grocery shop | 40 min | 55 min | 1.4 |
| Reply to the client brief | 20 min | 50 min | 2.5 |
| Tidy the kitchen | 15 min | 20 min | 1.3 |
After a week you will see a pattern, often a different one for different kinds of work. Writing tasks may run far over while chores are close. That ratio, from your own record, is a better correction than any rule of thumb.
Put the estimate to work
- Size the day with it. A realistic list depends on realistic durations; see how to write a realistic to-do list and how many to-dos you can really do in a day.
- Size time blocks with it. If you block time, block the corrected length; see time blocking on iPhone, the light way.
- Count back from deadlines with it. The do date for a big task comes from the corrected estimate, not the hopeful one; see deadline, due date or a day to do it.
When it still goes wrong
Some tasks resist estimates: things you have never done, work that waits on other people, anything involving a government office. For those, plan a halfway check rather than a better number. When half the time has gone, look at how much is done and decide then whether to move the deadline, cut the scope or ask for help. And if estimating starts to take longer than the tasks, stop overplanning is the page for you.
Where Dozen helps, and where it doesn’t
Now is a quiet screen with a timer for the one to-do you started, also on the Lock Screen and in the Dynamic Island. Starting a task there, and glancing at the clock when you finish, is an easy way to time a task once. For planning the day, Dozen counts finished to-dos rather than minutes, which avoids part of the guessing: Your usual day is the median number of to-dos you finished on your 14 most recent working days.
Where it doesn’t: if you want guesses and actual times side by side, a notebook or a spreadsheet is the simplest estimate log. And no app removes the bias itself; looking at your own past record is still the fix.
Questions and answers
Should I double every estimate?
A fixed rule like doubling is better than nothing, but your own ratio is better than a rule. Write down your guess and the real time for a week of tasks, divide one by the other, and use that number. It is often different for different kinds of work.
Why do other people seem better at estimating my tasks?
Because they do not have your plan in their head. In one of the original planning fallacy studies, observers predicting someone else’s task leaned on that person’s past record, and they erred on the long side rather than the short one.
Sources
What this page says about Dozen was checked against the app on .
- Buehler, Griffin and Ross (1994), Exploring the planning fallacy: why people underestimate their task completion times, Journal of Personality and Social Psychology 67(3) (University of Florida course copy) — thesis predictions of 33.9 days against 55.5 actual; people focus on future scenarios; connecting past experiences removed the bias in Study 4; observers overestimated others' times. Read 13 Nov 2026
- Wikipedia, Planning fallacy — Roy and colleagues' finding that memories of past task durations are also too short, and the segmentation effect. Read 13 Nov 2026
- Thomas and König (2018), Knowledge of previous tasks: task similarity influences bias in task duration predictions, Frontiers in Psychology 9 — predictions for a second task were less biased when the first task had been the same task. Read 13 Nov 2026
Get Dozen
Dozen is free to download. Dozen Pro, a monthly or yearly subscription, adds the Assistant.
Coming soon to iPhone and iPad
Dozen: Tasks & Notes
Dozen is free to download. Dozen Pro, a monthly or yearly subscription, adds the Assistant.
Coming soon to iPhone and iPad