The planning fallacy is the tendency to underestimate how long your own tasks will take, even when you know similar tasks ran late before. Researchers found that it shrinks when people deliberately connect their past experiences to the new estimate, which means the fix is a record of how long your work really takes, compared with what you planned. TimeFence from Automata Labs keeps that record for you: it shows an always-on-top timer while you work and saves each session’s planned duration and overtime to a private, exportable history.
See how long your work really takes, then plan the next block with real numbers.
Get it from Microsoft
What is the planning fallacy?
The term comes from Daniel Kahneman and Amos Tversky, who described it in 1979. In 1994, Roger Buehler, Dale Griffin and Michael Ross tested it in a well-known paper in the Journal of Personality and Social Psychology, “Exploring the ‘Planning Fallacy’: Why People Underestimate Their Task Completion Times”.
Their summary of Kahneman and Tversky’s idea is useful because it names the two kinds of information you can use when you estimate:
- Singular information “relates to aspects of the specific target task,” such as your plan for doing it.
- Distributional information “concerns how long it took to complete other, similar tasks.”
People who focus on the singular side “adopt an internal perspective: They concentrate on working out how they will complete the target task.” People who use the distributional side compare the new task with past ones. The planning fallacy is what happens when the plan wins and the history is ignored.
How big is the gap between planned and actual time?
Bigger than most people expect. In the paper’s first study, students predicted when they would finish their theses, and the thesis course coordinator recorded when each thesis was actually submitted.
| Measure from Study 1 | Result |
|---|---|
| Average best-estimate prediction | 33.9 days |
| Average actual completion time | 55.5 days |
| Share who finished by their best estimate | 29.7%, fewer than one third |
| Average “pessimistic” prediction | 48.6 days, still short of the actual time |
| Correlation between predicted and actual times | r = .77 |
Two details stand out. First, even the students’ worst-case predictions were too short. Second, the estimates were not random: students who expected to take longer did take longer. The authors put it this way: “Predictions can be informative even in the presence of a marked prediction bias.” In other words, your estimates contain real information. They are just shifted in one direction, and a shift is something you can measure and correct.
Across all of the paper’s prospective studies, the authors report that “fewer than one half of the participants finished their tasks in the amount of time they originally predicted.”
Why do we keep underestimating our own tasks?
The paper tested three explanations, and its abstract reports that “Results supported each hypothesis”:
- We underestimate our own tasks, but not other people’s. In one study, observers predicting someone else’s completion time actually overestimated it and “made greater use of relevant past experiences.”
- We think about the plan, not the past. When people were asked to think aloud while estimating, they “focused primarily on future scenarios.”
- We explain away past failures. People tended to blame earlier overruns on “relatively external, transient, and specific factors,” which makes those overruns feel irrelevant to the next task.
That third point matters for daily work. “Yesterday’s report took twice as long because of that meeting” may be true, but if a meeting interrupts most days, it is not a one-off. It is part of how long a report takes for you.
What actually reduces the planning fallacy?
The most useful finding in the paper is from Study 4. Simply remembering past assignments was not enough. But for the group asked to link those memories to the new task, the abstract reports that “the optimistic bias was eliminated for Ss instructed to connect relevant past experiences with their predictions.”
That gives you a practical rule: before you estimate, look at how long similar work took, and use that number on purpose. The hard part is that most people have no reliable record. Memory keeps the plan and edits out the overrun. A timer log does not.
| Approach | What it relies on | Typical result |
|---|---|---|
| Estimate from the plan | Singular information | Optimistic, as the research shows |
| Pad every estimate by a guess | A rule of thumb | Better, but not tied to your real pattern |
| Recall past tasks from memory | Memory, which tends to excuse overruns | Helps only if you link it to the new task |
| Check a log of planned vs actual time | Distributional information about you | Grounded in your own track record |
How do I track planned vs actual time on Windows?
You need three things: a planned duration for each block, a visible clock while you work so you notice when you go over, and a history you can read later. TimeFence covers all three. According to its Microsoft Store listing, it offers:
- Always-On-Top HUD: a visual progress ring that floats above your other windows while you work.
- Click-through overlay: you can “click, type, and scroll right through the timer overlay without it ever stealing your keyboard focus.”
- Phase chimes: audio cues when your sprint enters the Warning or Overtime phases, so going over the plan is something you notice rather than discover later.
- Session History and CSV Export: it “automatically track[s] your daily focus hours, planned durations, and overtime in a private, local dashboard,” and you can export the data.
- Strict Mode: it can disable the Pause and Quit buttons so you do not abandon a sprint early.
- Global hotkey: start or stop a 25-minute sprint from anywhere with Win+Alt+2.
- Offline and private: no telemetry, no cloud syncing and no account.
The overtime number is the important one here. Every time a block runs past the plan, TimeFence records how far it went, and that is exactly the distributional information the planning fallacy research says people leave out.
Turn every overrun into data, and let your history set the next estimate.
Get it from Microsoft
How do I turn the log into better estimates?
- Name your recurring task types. For example: weekly report, code review, client email batch, design revision. Keep the list short.
- Plan each block before you start. Set the planned duration you honestly expect, not the one you hope for.
- Let the timer run past the plan if needed. Do not stop at the chime just to make the numbers look good. The overtime is the data.
- Export the history after two or three weeks. Open the CSV and group sessions by task type.
- Work out your personal ratio. Divide total actual time by total planned time for each task type. A ratio of 1.4 means that kind of task usually takes 40% longer than you plan.
- Apply the ratio next time. Before you plan a new block, multiply your gut estimate by the ratio for that task type. This is the “connect past experience to the new task” step from the research, done with numbers instead of memory.
- Recheck monthly. Ratios change as you get faster, or as your work changes.
You do not need perfect data. Even a rough ratio built from a dozen sessions is better than a plan with no history behind it.
Does the planning fallacy affect people with ADHD or time blindness more?
The 1994 paper studied general student groups and does not address ADHD, so we will not claim a link it did not test. What we can say is practical: if you already find it hard to feel time passing, a visible timer and a written record remove two sources of error at once. You see the time while you work, and you do not have to rely on memory afterwards. Our guides on time blindness and desktop timers and focus timers for neurodivergent deep work go deeper on the first part.
If you want to compare TimeFence with built-in or free options, see Windows 11 Focus sessions vs TimeFence and TimeFence vs tray Pomodoro timers.
Frequently asked questions
Who discovered the planning fallacy?
Daniel Kahneman and Amos Tversky described it in 1979. Buehler, Griffin and Ross tested it in a series of studies published in 1994.
Is the planning fallacy the same as procrastination?
No. Procrastination is delaying the start. The planning fallacy is underestimating how long the task will take once you do it, and it happens even to people who start on time.
Does adding a buffer fix it?
A buffer helps, but a fixed guess is not tied to your real pattern. A ratio worked out from your own planned and actual times is more accurate.
Can TimeFence export my session history?
Yes. Its Store listing says it tracks focus hours, planned durations and overtime locally, with CSV export.
Does TimeFence need an account?
No. The Store listing says it works offline with no account and no telemetry.
Try TimeFence
Get TimeFence from the Microsoft Store. It is a one-time purchase with no subscription, and the Store listing shows the current price.
Keep time visible, log the overruns, and plan with the numbers next week.
Get it from Microsoft
Microsoft and Windows are trademarks of the Microsoft group of companies. TimeFence is not affiliated with or endorsed by the authors or publishers of the research cited.
Last updated 21 October 2026. Published by Automata Labs, a software company in Canada.