← Back to Paradoxes

The Planning Fallacy

Kahneman & Tversky's 1979 discovery: we systematically underestimate how long tasks take, even when we KNOW about this bias! The average project overruns by 25-50%. You're about to prove it on yourself.

🎯 Test Your Estimation Skills

Loading task...
I estimate this will take: hours

📊 The Reality

Your estimate:
Actual time:

0
Tasks Estimated
--%
Avg Underestimate
0
Times Underestimated
0
Times Overestimated
Task 1 of 8

🏗️ Famous Planning Disasters

Sydney Opera House
Est: 4 years, $7M Actual: 16 years, $102M
1,357% over budget
Denver Airport
Est: $1.7B, 1993 Actual: $4.8B, 1995
182% over budget, 16 months late
Big Dig (Boston)
Est: $2.6B, 1998 Actual: $14.6B, 2007
462% over budget, 9 years late
Eurofighter Typhoon
Est: £7B, 1997 Actual: £37B, 2003
429% over budget
Scottish Parliament
Est: £40M Actual: £431M
978% over budget
Channel Tunnel
Est: £5.5B Actual: £9.5B
80% over budget, 1 year late

🧠 Inside View vs Outside View

Inside View (What We Do)

We imagine the specific task, think through the steps, and estimate based on our unique situation. "This project is different—I know what I'm doing!"

Outside View (What Works)

Ask: "How long did SIMILAR projects take in the past?" Ignore the specifics, just look at base rates. This "reference class forecasting" is far more accurate.

The Paradox: Even when we KNOW about the planning fallacy, even when we've experienced it repeatedly, we STILL believe our current estimate is realistic. Each new project feels unique, so we take the inside view again.

🔬 The Research

Buehler et al. (1994): Students estimated their thesis completion at 33.9 days on average. Actual average: 55.5 days (64% overrun). Even their "worst case" estimates were exceeded by 20%!

Kahneman's Tax Form Study: Participants estimated 1 week to file. Average actual time: 3 weeks. Even those who remembered past overruns made optimistic predictions for the next task.

Flyvbjerg (2003): Analyzed 258 transportation projects worth $90B. Average cost overrun: 28%. Rail projects: 45%. 90% of projects exceeded estimates.

Why? We focus on the plan (what could go right) rather than obstacles (what could go wrong). We're optimistic about our abilities. We anchor on best-case scenarios. And we underweight "unknown unknowns."

✅ Debiasing Techniques

  • Reference Class Forecasting: Find similar past projects and use their actual times as your baseline.
  • Pre-mortem: Imagine the project failed. What went wrong? This surfaces hidden risks.
  • Multiply by π: Your estimate × 3.14. Sounds crazy, but surprisingly accurate for software projects.
  • Segment the task: Break into pieces, estimate each, then add. Sub-tasks are harder to be optimistic about.
  • Ask a friend: Others take the outside view naturally. They'll give more realistic estimates for YOUR work.
  • Track your record: Keep a log of estimates vs actuals. Calculate your personal "fudge factor" and apply it.