Cognitive Science: The Planning Fallacy
Why Your Brain Lies to You About How Long Things Take
The Sydney Opera House was supposed to take four years and cost a modest sum — it took fourteen years and cost roughly ten times the original estimate, and this was not a fluke.
The Idea
In 1979, Daniel Kahneman and Amos Tversky gave a name to something that had been quietly wrecking timelines since the first human said 'I'll have this done by Friday': the planning fallacy. The core finding is this — when people estimate how long a task will take, they systematically underestimate, even when they have direct personal experience of similar tasks going badly over time. This is not a gap in information. It is a gap in how we construct predictions. The mechanism is what makes this interesting. When we plan, we tend to imagine the specific future we intend — a clean, uninterrupted sequence of steps going roughly to plan. What we do not do is recall the base rate: how long this kind of project actually takes for people like us, in situations like this. Kahneman later called this the difference between the 'inside view' and the 'outside view'. The inside view feels vivid and personal and motivating. The outside view feels abstract and deflating. So we ignore it. There is also an optimism layer on top: we discount the possibility of unexpected obstacles, not because we think nothing will go wrong, but because we simply do not include disruptions in our mental simulation at all. The plan we imagine is a best-case scenario dressed up as a prediction. The result is that we are not bad at estimating tasks — we are bad at estimating the tasks we actually end up doing, which are always more complicated than the tasks we imagined when we started.
In the World
In the late 1990s, a team of psychologists ran a study at a Canadian university that has become one of the most cited demonstrations of the planning fallacy in action. Students were asked to estimate when they would finish their undergraduate thesis — and crucially, to give three estimates: a best case, a worst case, and what they realistically expected. Then the researchers just waited. The students, on average, predicted they would finish in about 34 days. Their own worst-case estimate was 49 days. The actual average? 55 days. More than half of them took longer than their own pessimistic scenario. When asked later, many reported they had known, at some level, that their thesis might go long — but they had optimised their formal prediction around the scenario they hoped for, not the one they had reason to expect. This pattern scales alarmingly. The Edinburgh tram project, Boston's Big Dig tunnel, the Berlin Brandenburg Airport — each began with confident, politically palatable estimates that bore almost no relationship to eventual costs and timelines. The Danish urban planner Bent Flyvbjerg spent decades cataloguing this in infrastructure projects worldwide and found consistent, massive overruns across cultures and continents. His conclusion was stark: the planning fallacy is not an error made by bad planners. It is the default output of human forecasting when left unexamined.
Why It Matters
Recognising the planning fallacy does not automatically fix it — Kahneman himself admitted he fell into it while writing about it — but understanding the mechanism suggests a practical correction. The most robust antidote researchers have found is 'reference class forecasting': instead of building your estimate from the inside out, start by asking how long things like this actually take for other people in comparable situations. Then adjust inward from there, rather than outward from your optimistic scenario. This reframe has consequences beyond scheduling. The planning fallacy shapes which projects we start, which commitments we make, how much buffer we leave in our lives. If your default forecast strips out friction, you will chronically overcommit — not because you are lazy or incompetent, but because the future you planned for was a fiction you mistook for a prediction. The deeper invitation here is to take the outside view more seriously in general: to ask, before assuming your situation is uniquely smooth, what the base rate actually is. It is one of the most transferable moves in all of decision science.
A Question to Ponder
Is there something you are currently planning — a project, a conversation, a change — where you are working from the best-case version of events, and what would the outside view actually say about it?
Get a new one of these every morning.
Start learning with Thinkable