The Planning Fallacy: An S-Tier Behavioral Designer’s Guide

In the 1970s, Daniel Kahneman assembled a team in Jerusalem to write a textbook on judgment and decision-making.

A year in, he asked everyone how long the rest would take. Every answer landed near two years.

Then he asked Seymour Fox, the curriculum expert in the room, how long teams like theirs had taken in the past.

About 40 percent never finished at all, Fox told him. None finished in under seven years.

Nobody quit. The textbook took eight years.

The researchers who discovered the planning fallacy could not out-think it at their own conference table.

Most write-ups stop at “add buffer time.” This guide goes further: why the bias survives your full awareness of it, why organizations quietly reward the worst estimates, and why I read the planning fallacy as a motivation problem wearing a forecasting costume.

Speed Run Notes

• The bias: people underestimate the time, cost, and risk of their own projects even when they know projects like theirs always run late. Kahneman and Tversky named it the planning fallacy in 1979.
• The signature stat: honors students predicted 33.9 days to finish their theses and took 55.5. Fewer than a third finished by their own estimate, and fewer than half by their worst-case date.
• The scale: across Bent Flyvbjerg’s 16,000-project database, 8.5% of big projects hit cost and time. Half a percent hit cost, time, and promised benefits. The Sydney Opera House ran ten years late.
• The mechanism: your brain answers “how long will this take” by simulating a success story instead of consulting the track record of similar projects. Kahneman called these the inside and outside views.
• The Octalysis read: an optimistic estimate is a reward you pay yourself at the planning table. Core Drive 2 collects the win now, Core Drive 8 buries the realistic number. Knowing better fixes neither.
• The fix: stop fighting motivation with motivation. Move the forecast outside the motivational field entirely: reference-class forecasting, task unpacking, and buffers designed into the system as features.

The terrain this guide covers, and where the depth sits:

• The origin: Kahneman’s eight-year textbook, the 1979 paper, and the inside view versus the outside view.
• The evidence: the honors thesis study, tax filers, the Sydney Opera House, and Flyvbjerg’s megaproject database.
• The verdict: what Kahneman and Tversky got right, plus three cracks in the framing — memory bias, strategic lying, and useful optimism.
• The science: constructive simulation, focalism, and why your brain builds highlight reels instead of forecasts.
• The comparisons: planning fallacy against optimism bias, Hofstadter’s Law, Parkinson’s Law, and hyperbolic discounting.
• The applications: megaprojects, AI-era software estimates, onboarding flows that weaponize the bias, and personal deadlines.
• The Octalysis merge: the Core Drive anatomy of an optimistic estimate, the designer’s ethical fork, and practical steps for planning around your own psychology.

Author Credibility: Yu-kai Chou

Yu-kai Chou created the Octalysis Framework after studying gamification since 2003 — years before the term entered mainstream vocabulary. As a Human-Systems Architect & Behavioral Designer, his framework has been applied by LEGO, Microsoft, Porsche, Coca-Cola, Salesforce, and MrBeast, impacting over 1.5 Billion Users.

Chou has taught the Octalysis methodology at Harvard, Stanford, Yale, Tesla, Google, BCG, and IDEO.

His work has been cited by Harvard, Stanford, MIT, Forbes, Wall Street Journal, Wired, US Department of Energy, NIST, NSF, NCBI, US Department of Education, ClinicalTrials.gov, and Google Scholar — with 3,700+ more academic publications. Explore his books here.

What Is the Planning Fallacy?

The planning fallacy is the systematic tendency to underestimate how long your own projects will take and how much they will cost, and it persists even when you know full well that similar projects always run over.

Daniel Kahneman and Amos Tversky named it in their 1979 paper “Intuitive Prediction: Biases and Corrective Procedures.”

The definition has a sharp edge that most summaries sand off.

The fallacy is measured against your own knowledge.

You have watched every product launch slip. You remember your last kitchen renovation.

You then estimate the next one as if none of that happened.

That combination is what makes it a fallacy rather than simple ignorance: the base rates are sitting in your head, unused.

Kahneman and Tversky traced it to a choice between two ways of forecasting.

The inside view builds the estimate from the project itself. You picture the steps, imagine the work going as planned, and add up the imagined durations.

The outside view ignores your project’s story entirely. It asks one question: how long did projects like this actually take?

Nearly everyone defaults to the inside view. The plan is vivid and available, and the base rate feels like it describes other people.

The result shows up in three reliable distortions: timelines come in too short, budgets come in too low, and benefits come in too high.

The third distortion gets the least press and does the most damage.

Flyvbjerg’s transport data found demand forecasts missing by wide margins in the optimistic direction, which means the business case was wrong on both ends: the project cost more than promised and delivered less than promised, and each error was sold as precision.

The Textbook That Took Eight Years

The founding story of the planning fallacy is Kahneman falling for it while writing a textbook about judgment errors, with the correct forecast sitting unused in the room.

The detail that matters is the double prediction.

When Kahneman asked his team for estimates, everyone answered from the inside view, including Seymour Fox, the curriculum dean who estimated around two years like the rest.

Only when Kahneman asked Fox a differently shaped question, how did teams like ours perform, did the outside view surface: roughly 40 percent of comparable teams never finished, and none finished in under seven years.

Fox held both numbers in his head at the same time and felt no contradiction.

That is the signature of the planning fallacy.

The base rate does not get argued down. It simply never gets invited into the forecast.

Worse, the team heard the outside view out loud, acknowledged it, and kept the two-year plan anyway.

Kahneman later wrote that the sensible move would have been to quit that day. Eight years later, the finished textbook arrived to a ministry that had lost interest in the project.

In the 1979 paper, Kahneman and Tversky formalized the distinction as singular versus distributional information: the story of this case versus the statistics of the class it belongs to.

Forty-seven years of research since has mostly confirmed a blunt rule. Forecasts built on singular information fail in predictable directions, and the class statistics were the better forecast all along.

The Evidence: From Honors Theses to 16,000 Megaprojects

The planning fallacy replicates at every scale researchers have checked, from student papers to national infrastructure, with error margins that should end careers and somehow never do.

The Honors Thesis Study

The cleanest demonstration is Roger Buehler, Dale Griffin, and Michael Ross’s 1994 study of psychology students finishing their honors theses.

Students predicted an average of 33.9 days. They took an average of 55.5 days.

Fewer than a third of them, 29.7 percent, finished by their own predicted date.

The researchers also asked for a worst-case estimate: how long would it take “if everything went as badly as it possibly could?”

The average worst-case answer was 48.6 days. Reality overshot their imagined catastrophe by a full week.

Fewer than half, 48.7 percent, finished even by that worst-case date.

These students could not construct a scenario pessimistic enough to contain their actual future, while their own history of late papers sat quietly in memory.

Tax Filers and Motivated Clocks

Buehler and colleagues later tracked Canadians predicting when they would mail their tax returns.

Filers predicted early submission and then filed well after their own estimates, and the people expecting refunds, the ones with money waiting, missed by more.

Wanting the outcome sooner made the prediction worse, which is an early clue that the fallacy runs on motivation rather than on missing information.

The Sydney Opera House

In 1957, the New South Wales government estimated its new opera house at 7 million Australian dollars, with doors open by January 1963.

The building opened in October 1973, ten years late, at a final cost of 102 million dollars.

That is a 1,357 percent overrun on the most famous silhouette in the southern hemisphere.

The architect, Jørn Utzon, had resigned mid-project and left the country. He never saw his finished building in person.

Flyvbjerg’s Database

Bent Flyvbjerg, the Oxford economic geographer who has spent his career measuring megaprojects, put numbers on the pattern at scale.

His 2002 study of 258 transportation projects found costs underestimated in roughly nine out of ten cases, with actual costs averaging 28 percent over estimate, and rail projects worst at around 45 percent.

His later database, built with Dan Gardner for the 2023 book How Big Things Get Done, now covers more than 16,000 projects.

The headline figures deserve a table.

Forecast survival rate, 16,000+ projects	Share of projects
Delivered on budget	Fewer than half
Delivered on budget and on time	8.5%
Delivered on budget, on time, and with promised benefits	0.5%

Flyvbjerg calls the pattern the iron law of megaprojects: over budget, over time, under benefits, over and over.

One caution on the numbers, since precision cuts both ways.

Completed-project databases can only count projects that finished, so the 40-percent-never-finish class from Fox’s outside view barely appears in them, a quiet survivorship bias that makes even these brutal statistics flattering.

What Kahneman and Tversky Got Right

Four of the pair’s original claims have held up for nearly five decades, and the most important one is that the fallacy is self-specific: we misforecast our own projects while judging other people’s just fine.

First, the self-other asymmetry.

Buehler’s team found that outside observers, given the same information about someone else’s task, produced far more accurate and more pessimistic completion estimates than the actors themselves.

Your colleague can see your deadline is fantasy. You can see theirs.

Neither of you can see your own.

Second, the persistence under awareness.

Telling people about the planning fallacy does not remove it. Students who reported their own past lateness in detail went on to predict optimistically anyway, treating history as a gallery of excuses rather than a forecast.

Each past overrun gets explained away by its circumstances, the flu, the flaky contractor, the surprise dependency, while the next project is imagined circumstance-free.

In hindsight every delay looks like an exception, which is hindsight bias working a second shift for the planning fallacy.

Third, the corrective actually corrects.

Their 1979 paper did more than diagnose; it prescribed. Identify the reference class, get its base-rate statistics, then adjust from that anchor only as far as your case demonstrably differs.

That procedure, now called reference-class forecasting, went from journal page to government policy: the UK Treasury’s Green Book guidance has required optimism-bias uplifts on public cost estimates since 2003.

When Kahneman was asked late in life which single piece of advice he would give forecasters, he answered with the outside view.

Fourth, the shape of the error is asymmetric.

Random errors would scatter in both directions. Planning errors point one way, toward optimism, which is what you would expect if the estimate were serving a psychological function rather than a statistical one.

Where the Planning Fallacy Framing Falls Apart

The standard cognitive-bias story has three cracks: some of the error is memory rather than prediction, some of it is lying rather than error, and some of it is load-bearing optimism we would not actually remove.

Maybe It’s Memory, Not Prediction

In 2005, Michael Roy, Nicholas Christenfeld, and Craig McKenzie argued in Psychological Bulletin that people underestimate future task duration partly because they misremember past task duration.

If your memory stores the kitchen renovation as three weeks when it took six, your next prediction can be perfectly calibrated to your records and still wildly wrong.

This account matters because it changes the fix. Debiasing the prediction does nothing if the corrupted input is the memory.

It also explains why keeping written time logs, boring as it sounds, outperforms most clever debiasing tricks: logs are the only version of your past that does not flatter you.

When It’s a Lie, Not a Bias

Flyvbjerg’s sharpest paper title asks the question directly: “Underestimating Costs in Public Works Projects: Error or Lie?”

His answer, for megaprojects, leans lie.

Project promoters compete for finite budgets, and the winning bid is reliably the prettiest forecast. Flyvbjerg calls this strategic misrepresentation, and he found the underestimation pattern too consistent over decades to be innocent error, since honest forecasters would learn.

A psychological bias and a political incentive produce the same lowball number, but they demand opposite treatments.

You cure a bias with better forecasting procedure. You cure strategic misrepresentation with accountability, audits, and consequences, because the forecaster already knows the truth.

Any serious discussion of the planning fallacy has to admit that in high-stakes settings, some share of the “fallacy” is a business model.

The Fallacy That Ships

The third crack is the uncomfortable one: the bias might be paying rent.

If founders saw true costs on day one, fewer companies would start. Kahneman himself acknowledged that optimistic overconfidence is the engine of capitalism, ruinous for the individual bettor and productive for the system that harvests the winners.

Martin Seligman’s work on learned optimism runs the same direction: optimistic explanatory styles predict persistence through setbacks, and persistence is precisely what long projects consume.

An honest account of the planning fallacy ends up two-handed.

The optimism that corrupts your schedule is often the same optimism that gets the thing attempted at all, so the design goal is containment rather than eradication: protect the forecast from the optimism without draining the optimism from the work.

That distinction, optimism about capability versus honesty about arithmetic, is the one the Octalysis section below turns into a design method.

What’s Really Happening Inside the Brain

Your brain does not retrieve the future the way it retrieves a fact; it constructs the future the way it constructs a story, and the construction engine is biased toward clean drafts.

Memory researchers Daniel Schacter and Donna Rose Addis describe remembering and imagining as one shared system: the same network that reconstructs your past assembles your simulated future from recombined pieces of memory.

That architecture has a known failure mode. Simulations run schematic and smooth.

When you imagine writing the report, the simulation renders you writing the report. It does not render the sick child, the corrupted file, or the reorg, because none of those belong to the concept “writing the report.”

Psychologists call the narrowing focalism: the plan fills the frame, and everything outside the plan, which is where delays actually live, goes unrendered.

Ian Newby-Clark and colleagues showed how deep the narrowing runs in 2000. When they asked people for realistic predictions and best-case predictions, the two answers came out nearly identical.

Your “realistic” estimate is your fantasy with a straight face. The simulation engine cannot tell them apart, and it hands both to you stamped with the same confidence.

Construal level theory adds the distance problem: events months away get simulated abstractly, as goals and meanings, while events today get simulated concretely, as steps and obstacles.

Distant projects therefore feel frictionless in a way tomorrow’s errands never do, and Gal Zauberman and John Lynch found we expect our future selves to have far more slack time than our present selves ever report having.

Add the Roy memory distortion from the last section, compressed recollections of how long things took, and the estimating pipeline is corrupted at every stage: flattering inputs, schematic simulation, and a confidence stamp at the end.

One more gear turns underneath: fluency.

A simulation that runs smoothly feels true, because the mind reads its own ease of processing as evidence about the world.

Your project plan is the single most rehearsed story you own. You have pitched it, defended it, and refined it, and every rehearsal makes the success path render faster.

By the time you estimate, the plan plays like a memory of something that already happened, and memories of things that happened do not feel like guesses.

None of that pipeline is a reasoning error you can argue with. It is rendering architecture, which is why the working fixes all route around the renderer instead of upgrading it.

Planning Fallacy vs Other Theories

The planning fallacy sits in a family of time-and-optimism effects, and confusing it with its relatives leads designers to prescribe the wrong intervention.

Planning Fallacy vs Optimism Bias

Optimism bias is dispositional and general: most people rate their odds of divorce, disease, and disaster as below average.

The planning fallacy is task-specific and sneakier, because even people who are pessimistic about life produce sunny schedules for their own projects.

The difference is mechanism. Optimism bias is a belief about outcomes, while the planning fallacy is a forecasting procedure error, the inside view, that manufactures optimistic outputs from neutral machinery.

That is why debiasing differs: you can talk someone down from an optimistic belief, but you cannot talk them out of a simulation. You have to replace the procedure.

Planning Fallacy vs Hofstadter’s Law

Douglas Hofstadter compressed the whole literature into one recursive joke in 1979, the same year as Kahneman and Tversky’s paper: “It always takes longer than you expect, even when you take into account Hofstadter’s Law.”

The joke is a real empirical claim. Buehler’s students who knew their history still predicted optimistically, and correction attempts routinely undershoot.

Hofstadter’s Law is the planning fallacy’s persistence clause: awareness of the bias becomes one more input the inside view cheerfully ignores.

Planning Fallacy vs Parkinson’s Law

Parkinson’s Law, work expands to fill the time available, sounds like the opposite claim, and managers regularly weaponize it to justify aggressive deadlines.

Both laws are true, and they operate in sequence.

The planning fallacy sets an impossible baseline, and Parkinson’s Law guarantees that whatever padding survives gets absorbed anyway.

Eliyahu Goldratt added the third gear in Critical Chain: student syndrome, the tendency to start work at the last moment the deadline allows.

Grant a buffer and it gets consumed before the task begins, which is why naive padding fails as a planning-fallacy cure.

The practical synthesis: tight deadlines do compress work, and they compress it by an amount your optimistic estimate already spent twice. Buffers work only when they are owned by the plan, pooled and visible, instead of donated silently to each task.

Planning Fallacy vs Hyperbolic Discounting

The planning fallacy misprices duration, while hyperbolic discounting misprices the pain of that duration, shrinking distant costs until they feel weightless from here.

Piers Steel’s temporal motivation theory then predicts the behavioral consequence: motivation stays flat until the deadline looms, so the underestimated schedule gets executed almost entirely in its final stretch.

Chain the three and you get the standard project tragedy: promise it fast, feel no urgency, sprint at the end, deliver late, remember it fondly, repeat.

The Planning Fallacy in the Real World

The same bias that delays a thesis by three weeks delays a railway by a decade, and in 2026 it has found a brand-new accelerant: AI-assisted work that feels faster than it measures.

Megaprojects and Government

Reference-class forecasting is the planning fallacy’s one genuine policy victory.

Since 2003, the UK Treasury’s Green Book supplementary guidance has required optimism-bias uplifts: mandatory percentage additions to cost estimates, sized by project type, applied before approval.

The logic is pure outside view. The government stopped asking whether this project’s estimate is honest and started pricing the historical error of its reference class directly into the budget.

Flyvbjerg’s consultancy work extended the method to rail, road, and building programs in multiple countries, and his data shows the uplift ranges are rarely flattering: some project classes need buffers north of 50 percent before the odds turn fair.

Software and the AI Era

Software estimation was already the planning fallacy’s home turf, and AI just handed the inside view a megaphone.

In July 2025, the research group METR published a randomized controlled trial that should be pinned above every sprint board.

Sixteen experienced open-source developers completed 246 real tasks on mature repositories, randomized between AI-assisted and AI-free conditions.

With AI tools, they took 19 percent longer.

Before starting, they expected AI to make them 24 percent faster.

Afterward, having lived the slowdown, they still believed it had made them 20 percent faster.

METR frames the result as a snapshot of early-2025 tools and workflows, and later tools may deliver.

The durable finding is the gap itself: felt speed and clocked speed pointed in opposite directions, which is the planning fallacy’s mechanism caught on camera.

“AI will make this fast” is 2026’s inside view. It is vivid, it is plausible, and it substitutes a simulation for a measurement.

The teams getting real value from AI run the Lean Startup move on their own tooling: measure cycle time before and after adoption, and let the clock outvote the vibes.

Roadmaps, OKRs, and the Quarter That Plans Itself Broke

Quarterly planning rituals give the planning fallacy a recurring stage with a captive audience.

A roadmap is a portfolio of inside views, one per initiative, each estimated by the team that proposed it and wants it approved.

Stack ten optimistic estimates and the errors do not cancel; they compound, because every initiative was priced without the interruptions the other nine will generate.

This is how organizations end each quarter having shipped 60 percent of the commitments while every individual team swears its own estimate was reasonable.

Peter Drucker’s Management by Objectives, the ancestor of the modern OKR, actually anticipated the trap: Drucker insisted objectives be grounded in evidence and negotiated against capacity, never declared from ambition.

The modern fix is portfolio-level honesty. Cap planned work below measured capacity, price cross-team dependencies explicitly, and grade last quarter’s forecast accuracy before anyone pitches the next one.

Teams that publish their own predicted-versus-actual history before planning day inherit the outside view for free, because the reference class is sitting in the room’s shared memory instead of each planner’s private optimism.

Product Design: The Weaponized Fallacy

Here is the section most bias write-ups will not print: product designers do not suffer from users’ planning fallacy. They monetize it.

“Setup takes 2 minutes.” “A quick 3-question survey.” “Cancel anytime in seconds.”

Every one of those promises is an inside view professionally manufactured for you, because low predicted effort is what gets you in the door.

The Fogg Behavior Model makes the mechanics explicit: behavior fires when motivation, ability, and prompt converge, and perceived ability is what the “2 minutes” claim inflates.

Once you are inside, the sunk minutes do the retention work.

I wrote about that trap as the Sunk Cost Prison: the design pattern where invested time, real or merely felt, becomes the wall that keeps players grinding.

The planning fallacy is the prison’s front gate. You walked in because the visit looked short.

Personal Productivity

At the personal scale, the two interventions with the strongest evidence are unpacking and implementation intentions.

Justin Kruger and Matt Evans showed in 2004 that forcing people to enumerate a task’s steps before estimating meaningfully reduces the fallacy, because unpacking drags obstacles into the simulation’s frame.

Peter Gollwitzer’s implementation intentions attack the execution side: pre-deciding “if situation X, then action Y” roughly doubled goal completion rates across a large meta-analytic literature, with effects in the medium-to-large range.

Neither fixes the estimate directly. Unpacking feeds the renderer better inputs, and if-then plans make the underestimated schedule less fatal by cutting execution drag.

The honest personal fix for the estimate itself remains the boring one: your own logged history, consulted like a stranger’s.

The Elephant in the Room

We have known the cure since 1979, it costs almost nothing, and nearly nobody uses it, because organizations do not actually want accurate forecasts. They want fundable ones.

The incentive structure explains the whole puzzle.

Two teams pitch for the same budget.

Team A took the outside view and quoted fourteen months. Team B quoted six, from the heart.

Team B gets funded.

Fourteen months later, Team B is over deadline and asking for more money, and the sunk investment argues for finishing rather than canceling.

Flyvbjerg calls the equilibrium survival of the unfittest: the projects with the most distorted forecasts are the ones most likely to win approval, so the system selects for the deepest bias.

Inside companies the selection is softer and identical. The optimistic estimate reads as confidence and commitment, while the accurate one reads as sandbagging, and everyone in the room can feel which answer the boss’s face is requesting.

This is why forty-seven years of debiasing advice have barely moved the megaproject statistics.

The planning fallacy persists at the civilizational scale because it keeps winning the game we make forecasts inside of.

Any fix that ignores the game, and coaches the individual, is treating a market failure with a mindfulness app.

Applying the Planning Fallacy with the Octalysis Framework

Run the planning fallacy through the Octalysis Framework and the bias stops looking like a forecasting error, because an optimistic estimate is a motivational payment: the plan is the first reward the project ever pays you.

Standard debiasing advice treats the estimate as a calculation that went wrong and prescribes better arithmetic.

Two decades of behavioral design have taught me to ask a different question about any stubbornly irrational behavior: what is it paying the person who does it?

An optimistic estimate pays out on at least five Core Drives at once, which is why it survives everything thrown at it.

The Core Drive Anatomy of an Optimistic Estimate

Core Drive 2 (CD2): Development & Accomplishment collects first.

A finished plan with a tight timeline feels like progress before any work exists. Planning is the one phase of a project where the win-state is purchasable by imagination alone, and the shorter the schedule, the bigger that imaginary win feels.

Core Drive 8 (CD8): Loss & Avoidance polices the other direction.

Quoting the honest number means losing something today: the stakeholder’s smile, the project’s approval, your self-image as the fast one. The overrun is also a loss, but it is distant, diffuse, and probably attributable to circumstances.

An immediate concrete loss versus a distant deniable one is no contest, and hyperbolic discounting has already rigged the exchange rate.

Core Drive 5 (CD5): Social Influence & Relatedness turns the estimate into a performance.

In any room where estimates compete, the optimistic one signals commitment and the accurate one signals doubt, so the forecast stops being a prediction and becomes a bid for standing.

Core Drive 1 (CD1): Epic Meaning & Calling supplies the moral cover.

On mission-driven teams, the schedule and the cause fuse, and questioning the timeline reads as questioning the mission. The more meaningful the project, the more protected its worst numbers become.

Core Drive 4 (CD4): Ownership & Possession seals it.

It is my plan. The endowment effect prices owned things above identical unowned things, and an owned schedule resists correction the way an owned house resists an honest appraisal.

Core Drive	What the optimistic estimate pays	When it collects
CD2: Development & Accomplishment	The win-state of a finished, impressive plan	Immediately, at the planning table
CD8: Loss & Avoidance	Escape from the immediate loss of quoting reality	Immediately
CD5: Social Influence & Relatedness	Standing, in rooms where estimates compete	Immediately
CD1: Epic Meaning & Calling	Loyalty to the mission the schedule represents	Immediately
CD4: Ownership & Possession	Pride of authorship in the plan itself	Immediately
The overrun	Every cost of the above	Months later, payable by future-you

Notice the column that matters: every reward lands now, and the entire cost lands on a person you have never met, future-you, whom Zauberman’s research shows you believe to be rich in free time.

This is why I call the planning fallacy a motivation problem wearing a forecasting costume.

Willpower-based fixes fail because they ask motivation to defeat motivation with the same currency. Structural fixes work because they move the forecast outside the motivational field entirely, which is exactly what reference-class forecasting does: nobody’s Core Drives are invested in a base rate.

Priced in Discovery, Paid in Scaffolding

Octalysis divides every user journey into four Experience Phases: Discovery, Onboarding, Scaffolding, and the Endgame.

The planning fallacy is a currency-exchange scam between two of them.

Estimates get priced in Discovery, the phase where the project exists only as imagination and every Core Drive pays out in simulated coin: imagined mastery, imagined applause, imagined shipping day.

The work gets paid for in Scaffolding, the long middle phase where motivation must come from real progress against real friction, at real exchange rates.

Discovery currency always overvalues. Scaffolding always collects in full.

Design teams that re-estimate at the Discovery-to-Scaffolding border, one sprint in, once real friction has repriced the work, catch the overrun while it is still a forecast instead of a confession.

White Hat Planning, Black Hat Deadlines

Octalysis divides motivation into White Hat drives, the top of the octagon, which build a sense of meaning, growth, and control, and Black Hat drives, the bottom, which create urgency through fear, scarcity, and loss.

The planning fallacy books the project’s motivation in the wrong hat.

Planning phase runs pure White Hat: growth simulated, meaning affirmed, control felt. Execution then inherits a schedule only Black Hat pressure can rescue, so the project’s back half is administered through deadline panic, and the panic gets remembered as “how shipping feels.”

Well-designed goal-setting spends the hats in the opposite order: honest, evidence-anchored timelines up front, so the closing sprint can be a chosen intensity rather than a structural apology.

The Designer’s Fork

If you design products, you face the planning fallacy twice, and the second encounter is an ethics exam.

Your users run the same biased simulation your project team does. They believe the setup is quick, the course finishable by Friday, the subscription cancelable the week they stop using it.

You can exploit that, and the industry mostly does: underquote the effort, collect the signup, let sunk cost handle retention.

Or you can design for the actual completion curve instead of the imagined one.

Games, at their best, already do.

A streak freeze in a language app is the planning fallacy priced into the design: the system knows you will miss a day before you believe it, and it builds forgiveness in advance.

Save points, grace periods, and catch-up mechanics are all buffers shipped as features instead of shame.

My rule for the fork: design for predicted effort and you convert a signup; design for actual effort and you keep a human. The first is a transaction, the second is a relationship, and the second is where every durable product I have advised makes its money.

Practical Steps: Planning Around Your Own Psychology

You cannot debias your simulation, so route around it: anchor on history, unpack the work, pre-decide the obstacles, and split the dreamer from the forecaster.

Each step below moves some part of the forecast out of the motivational field and into the evidence, which is the only trade the planning fallacy reliably loses.

1. Find your reference class before you estimate. List three to five completed projects that resemble this one, take their actual durations, and let that range be your anchor. Adjust away from it only for differences you can name out loud.
2. Compute your personal overrun multiplier. Divide actual time by predicted time for your last few tracked projects. Most people find a stable ratio between 1.4 and 2.5. Multiply every future gut estimate by yours, mechanically, no exceptions granted by enthusiasm.
3. Unpack before you quote. Enumerate the steps in writing first; per Kruger and Evans, decomposition drags obstacles into view that the smooth simulation omits. Estimate the pieces, then add integration time on top.
4. Write if-then plans for the first three obstacles. “If the API access is delayed, then I start on the data model.” Implementation intentions cut execution drag even when the estimate stays wrong.
5. Separate the wanting from the forecasting. Whoever champions the project should never be the sole author of its timeline. Give the forecast to someone whose Core Drives are not invested in the answer, the way the Green Book gives it to a base rate.
6. Log actuals, not impressions. Keep a plain record of predicted versus actual on everything meaningful. Your memory will compress past durations; the log is the only witness that does not flatter you.
7. Ship buffers as features. Whether the user is your team or your customer, build the grace period into the system: schedule slack owned by the plan, streak freezes, catch-up paths. Forgiveness designed in advance beats apology issued after.

The Planning Fallacy Was the Beginning, Not the End

Kahneman and Tversky’s 1979 paper aged into two different legacies: a psychology of how minds simulate the future, and an engineering discipline for making promises that survive contact with it.

The psychology line ran through Buehler’s field studies, the memory-bias challenge, and today’s prospection research, which reframed the fallacy as the signature of a simulation engine that builds futures from flattering parts.

The engineering line ran through Flyvbjerg into national policy, where the outside view stopped being advice and became a spreadsheet cell with legal standing.

Behavioral design is where the two lines meet, because every product roadmap, onboarding flow, and habit loop is a promise about someone’s future time.

I wrote Actionable Gamification around a conviction that fits this bias exactly: systems built for how humans actually behave beat systems built for how humans are supposed to behave.

Your move this week is small and concrete.

Pull your last three finished projects, divide actual by predicted, and write the multiplier somewhere your next planning meeting cannot avoid seeing it.

Then watch the room negotiate with your own history, and notice which Core Drives do the talking.

Frequently Asked Questions

What is the planning fallacy in simple terms?

The planning fallacy is the tendency to underestimate how long your own tasks will take and how much they will cost, even when you know that similar tasks have always run over. Your knowledge of past overruns simply does not get used when you forecast your own next project.

Who coined the term “planning fallacy”?

Daniel Kahneman and Amos Tversky, in their 1979 paper “Intuitive Prediction: Biases and Corrective Procedures.” The same paper proposed the cure that later became reference-class forecasting.

What is the difference between the planning fallacy and optimism bias?

Optimism bias is a general belief that bad outcomes are less likely for you. The planning fallacy is a forecasting procedure error specific to your own tasks: even pessimists produce optimistic schedules for their own projects, because they estimate from an imagined success story rather than from base rates.

What causes the planning fallacy?

The core cause is the inside view: people forecast by mentally simulating their project going as planned, and the simulation omits obstacles by design. Compressed memories of past durations, focus on the plan instead of its surroundings, and the social rewards of confident estimates all deepen the error.

What is a famous example of the planning fallacy?

The Sydney Opera House. Estimated in 1957 at 7 million Australian dollars with completion targeted for January 1963, it opened in October 1973 at 102 million dollars, ten years late and roughly 1,357 percent over budget.

How do you overcome the planning fallacy?

Use the outside view: anchor estimates on the actual durations of similar completed projects, apply your personal or institutional overrun ratio, unpack tasks into steps before estimating, and separate whoever wants the project from whoever forecasts it.

Awareness alone does not work.

What is reference-class forecasting?

Reference-class forecasting estimates a project by consulting the statistical track record of the class of similar past projects instead of the project’s own plan. The UK Treasury’s Green Book has required this style of optimism-bias uplift on public cost estimates since 2003.

Is the planning fallacy ever useful?

Sometimes. Optimistic forecasts get hard projects attempted, funded, and persisted with, and Kahneman acknowledged optimism as an engine of economic dynamism.

The design goal is containment: keep the optimism in the work while keeping it out of the arithmetic.

Does AI fix the planning fallacy?

Not so far, and it may amplify it. In METR’s 2025 randomized trial, experienced developers using early-2025 AI tools took 19 percent longer on real tasks while believing AI had made them about 20 percent faster.

Expected speedups from new tools are inside-view estimates and deserve the same outside-view discipline as any other forecast.

How does the Octalysis Framework explain the planning fallacy?

Octalysis reads the optimistic estimate as a motivational payment: Core Drive 2 collects an instant sense of accomplishment, Core Drive 8 avoids the immediate loss of quoting reality, and Core Drive 5 wins standing in rooms where estimates compete. The costs all land on future-you, which is why structural fixes beat willpower.

References

• Kahneman, D., & Tversky, A. (1979). Intuitive prediction: Biases and corrective procedures. TIMS Studies in Management Science, 12, 313-327.
• Buehler, R., Griffin, D., & Ross, M. (1994). Exploring the “planning fallacy”: Why people underestimate their task completion times. Journal of Personality and Social Psychology, 67(3), 366-381.
• Buehler, R., Griffin, D., & MacDonald, H. (1997). The role of motivated reasoning in optimistic time predictions. Personality and Social Psychology Bulletin, 23(3), 238-247.
• Newby-Clark, I. R., Ross, M., Buehler, R., Koehler, D. J., & Griffin, D. (2000). People focus on optimistic scenarios and disregard pessimistic scenarios while predicting task completion times. Journal of Experimental Psychology: Applied, 6(3), 171-182.
• Flyvbjerg, B., Holm, M. S., & Buhl, S. (2002). Underestimating costs in public works projects: Error or lie? Journal of the American Planning Association, 68(3), 279-295.
• Kruger, J., & Evans, M. (2004). If you don’t want to be late, enumerate: Unpacking reduces the planning fallacy. Journal of Experimental Social Psychology, 40(5), 586-598.
• Roy, M. M., Christenfeld, N. J. S., & McKenzie, C. R. M. (2005). Underestimating the duration of future events: Memory incorrectly used or memory bias? Psychological Bulletin, 131(5), 738-756.
• Zauberman, G., & Lynch, J. G. (2005). Resource slack and propensity to discount delayed investments of time versus money. Journal of Experimental Psychology: General, 134(1), 23-37.
• Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. Advances in Experimental Social Psychology, 38, 69-119.
• Schacter, D. L., & Addis, D. R. (2007). The cognitive neuroscience of constructive memory: Remembering the past and imagining the future. Philosophical Transactions of the Royal Society B, 362(1481), 773-786.
• Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux. (Chapter 23, “The Outside View.”)
• HM Treasury (2003, updated). The Green Book: Supplementary Guidance on Optimism Bias. UK Government.
• Flyvbjerg, B., & Gardner, D. (2023). How Big Things Get Done. Currency / Random House.
• Becker, N., et al. / METR (2025). Measuring the impact of early-2025 AI on experienced open-source developer productivity. arXiv:2507.09089.
• Hofstadter, D. (1979). Gödel, Escher, Bach: An Eternal Golden Braid. Basic Books.

Related Reading

This pillar is part of the Behavioral Framework Library. If the planning fallacy earns a bookmark, these neighbors extend it:

• Confirmation Bias: How We Select the Evidence That Agrees With Us
• Negativity Bias: Why Bad Is Stronger Than Good
• The Goal Gradient Hypothesis: Why Motivation Accelerates Near the Finish
• The Fresh Start Effect: Temporal Landmarks and New Beginnings
• The Zeigarnik Effect: Why Unfinished Tasks Occupy Your Mind

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?