Planning Fallacy¶

Prime #: 1066
Origin domain: Cognition And Psychology
Subdomain: judgment and decision making → Cognition And Psychology

Core Idea¶

The planning fallacy is the structural pattern in which agents forecasting the time, cost, or risk of executing a novel task systematically underestimate those quantities — even when they know that similar past tasks have run long, over budget, or hit unanticipated failures. The bias is robust to feedback, persists in experts, and has been measured with roughly the same magnitude — a median overrun of about 30–50% on time and cost, with a long right tail — across decades, cultures, and project classes.

The load-bearing mechanism is inside-view substitution for the outside view. When constructing a forecast, the planner mentally simulates the steps they intend to take, the obstacles they anticipate, and the resources they plan to deploy, and produces an estimate consistent with that simulation. This inside view omits the long tail of unanticipated interruptions, partial failures, dependency surprises, and known-but-unmodelled overhead — because, by construction, the simulation contains only the steps the planner can imagine. The outside view — the empirical distribution of completion times for the reference class of comparable past tasks — is far more informative, but it is psychologically unavailable to a planner committed to the inside-view simulation ("but my project is different"), and so goes unused. The structural commitment is that two distinct forecasting strategies — simulate the plan and reference the empirical distribution — produce systematically different estimates, and that planners default to the first, which is biased downward. Crucially, the bias is produced by an otherwise reasonable process operating on an unrepresentative reference class — the set of actions one can imagine, which omits the unimaginable interruptions that will actually occur — so it is a feature of the strategy's blind spot, not of forecaster carelessness, which is why effort and expertise do not remove it.

How would you explain it like I'm…

The Five-Minute Fib

When you start coloring a big picture, you think, 'I'll be done in five minutes!' But then a crayon breaks, the dog wants to play, and dinner is ready, so it takes way longer. The Planning Fallacy is how we almost always guess things will be quicker and easier than they really turn out. We forget about all the little surprises that pop up.

Always Takes Longer

The Planning Fallacy is the way people guess that a new task will take less time and money than it really does, even when they know past projects ran long. The reason is that when you plan, you picture the steps you mean to take, and that picture looks smooth and quick. But it leaves out the traffic jams, the broken tools, and the waiting on other people, because you can't imagine surprises you've never thought of. A better trick is to look at how long similar jobs actually took before, instead of trusting your mental movie. People skip that trick because they feel 'my project is different.' That's why even experts who plan a lot still fall for it.

Inside-View Underestimate

The Planning Fallacy is the robust pattern in which people forecasting the time, cost, or risk of a novel task systematically underestimate it, even knowing that similar past tasks ran long or over budget. The mechanism is using the inside view instead of the outside view: when you forecast, you mentally simulate the steps you plan to take and the obstacles you foresee, and your estimate matches that simulation. But the simulation only contains steps you can imagine, so it omits the long tail of unanticipated interruptions, partial failures, and dependency surprises. The outside view (the actual spread of completion times for a reference class of comparable past tasks) is far more accurate, but a planner committed to 'my project is different' rarely reaches for it. So two strategies, simulate-the-plan versus check-the-distribution, give systematically different answers, and we default to the first, which is biased low. Because the bias comes from an otherwise reasonable process working on an incomplete picture, more effort and expertise do not remove it.

The planning fallacy is the structural pattern in which agents forecasting the time, cost, or risk of executing a novel task systematically underestimate those quantities, even when they know that similar past tasks ran long, over budget, or hit unanticipated failures. It is robust to feedback, persists in experts, and shows roughly constant magnitude across decades and cultures: a median overrun of about 30 to 50 percent on time and cost, with a long right tail. The load-bearing mechanism is inside-view substitution for the outside view. Constructing a forecast, the planner mentally simulates the intended steps, the anticipated obstacles, and the planned resources, and produces an estimate consistent with that simulation. The inside view omits the long tail of unanticipated interruptions, partial failures, dependency surprises, and known-but-unmodelled overhead, because by construction the simulation contains only steps the planner can imagine. The outside view, the empirical distribution of completion times for the reference class of comparable tasks, is far more informative, but it is psychologically unavailable to a planner committed to the inside-view simulation ('but my project is different'). The structural commitment is that two distinct strategies, simulate-the-plan and reference-the-empirical-distribution, yield systematically different estimates, and planners default to the first, which is biased downward. The bias is produced by an otherwise reasonable process operating on an unrepresentative reference class (the actions one can imagine, which omit the interruptions that will actually occur), so it is a property of the strategy's blind spot, not of carelessness, which is why effort and expertise do not remove it.

Structural Signature¶

the forecasting agent facing a novel goal-directed task — the inside-view simulation strategy — the outside-view reference-class distribution — the default substitution of the first for the second — the simulation's structural blind spot (the omitted long tail) — the downward-biased, feedback-resistant estimate

The pattern is present when each of the following holds:

A forecasting agent. Some planner must estimate the time, cost, or risk of executing a task whose particulars are at least partly novel to it.
Two distinct strategies. Two well-defined forecasting strategies are available: inside-view simulation of the intended steps and obstacles, and outside-view reference to the empirical distribution of comparable past attempts.
A default substitution. The agent defaults to the inside view — the vivid, self-coherent simulation — and the reference class, though available, goes psychologically unused ("my case is different").
A structural blind spot. The simulation can only contain steps the planner can imagine, so it omits the long tail of unanticipated interruptions and dependency surprises that the reference class captures empirically.
A directional bias. The substitution skews the estimate systematically downward — a robust median overrun with a long right tail — rather than producing symmetric error.
Feedback resistance. Because the cause is the strategy's blind spot rather than carelessness or missing information, the bias survives expertise, effort, and explicit feedback; finer inside-view detail worsens it.

These compose into a non-obvious commitment: unbiased forecasts of intentional action come from the empirical distribution of comparable attempts, not from internal simulation of the path, so the corrective is structural — force the outside view into the estimate — not informational.

What It Is Not¶

Not optimism_bias. Optimism bias is a general directional skew toward favorable outcomes across all judgments. The planning fallacy is a specific mechanism — inside-view simulation substituting for the outside-view reference class — that produces underestimation in forecasts of one's own task. The fallacy explains why the optimism appears in planning; it is not the bias in general (see optimism_bias).
Not strategic misrepresentation. When a promoter deliberately low-balls a forecast to win approval, the forecaster is not mistaken — the cause is incentive, not a cognitive blind spot. The planning fallacy is strictly the case where the forecaster genuinely believes the low estimate. The discriminator: does the forecaster bear the overrun's cost?
Not scenario_planning. Scenario planning is a remedy-family tool — imagining multiple futures, including failures. The fallacy is the disease: the default to a single self-coherent simulation. Scenario planning and pre-mortems are outside-view correctives, not instances of the fallacy itself.
Not bounded_rationality. Bounded rationality is the general limit on cognitive resources. The planning fallacy is feedback-resistant even with unlimited deliberation: more effort and finer decomposition worsen it. It is a blind spot of a particular strategy, not a resource shortfall.
Not overfitting. Both involve a model that fits known cases yet generalizes poorly, but overfitting is a statistical property of a fitted model on data; the planning fallacy is a forecasting-strategy default in an agent who has a reference class but won't use it.
Common misclassification. Prescribing "plan more carefully" or finer task decomposition as the fix. Because the bias lives in the inside-view strategy's blind spot, deeper simulation deepens the bias. The tell: if the proposed remedy adds inside-view detail rather than importing the empirical distribution, it will make the estimate more confident and more wrong.

Broad Use¶

The same inside-view substitution recurs across every domain where agents plan novel goal-directed action. In megaprojects — transport, energy, infrastructure — decades of systematic time and cost overruns show a characteristic right-skewed distribution that does not improve over time despite repeated promises of better planning. In software engineering, nearly every project of non-trivial scope ships late; the discipline's estimation tools — story points, velocity tracking, three-point estimation — are functionally outside-view scaffolds, and "small batches" practices are partly a response to the structural difficulty of inside-view estimation at any meaningful horizon. In personal and academic work, the original thesis-completion studies found students' best-, most-likely-, and even worst-case estimates all routinely exceeded. In surgery and clinical scheduling, procedures run long, and operating rooms have moved to empirically calibrated per-surgeon, per-procedure time estimates precisely because inside-view estimates were biased low. In military planning, "no plan survives contact with the enemy" is the soldier's restatement, met by branch-and-sequel analysis, reserves, and pre-mortems — all outside-view tools. In emergency and pandemic planning, inside-view scenarios underestimate long-tail demand, and reference-class forecasting against historical events improves capacity sizing. In everyday tasks — cooking, home renovation — the underestimation is ubiquitous enough to be folklore. In each, the substrate is "a forecaster with access to a relevant reference class but defaulting to inside-view simulation," and the same structural signature appears: right-skewed overrun, robust median magnitude, non-elimination by feedback, and a structural rather than informational fix.

Clarity¶

The fallacy clarifies by naming the dissociation between two well-defined forecasting strategies — inside-view simulation and outside-view reference-class — and locating the bias not in either strategy alone but in the default substitution of one for the other. This separates three errors practitioners frequently merge: that the planner lacks information about the task (often false — they usually have it); that the planner is dishonest about the estimate (sometimes true but not necessary for the bias); and that the planner has the information but is using the wrong forecasting strategy (the load-bearing claim). It also names a striking non-learning asymmetry: unlike most biases, the planning fallacy is not eliminated by explicit feedback. A planner can agree that past projects ran long, agree that this project is comparable, and then estimate again from the inside view and produce another low number. The frame makes the fix legible as structural — switch the forecasting strategy — rather than informational — supply more facts about the task.

Manages Complexity¶

The fallacy compresses a vast practitioner literature on "why projects are late" into one diagnostic schema: identify which forecasting strategy produced the estimate; ask whether the outside-view distribution would predict a different number; apply reference-class adjustment or an empirical buffer. The schema generalizes across substrates without modification — the same diagnostic applies to a subway project, a software release, a thesis chapter, and dinner. The compression also captures a class of interventions that do not work: exhortations to plan more carefully, finer-grained Gantt charts, and deeper task decomposition typically worsen the bias by deepening the inside-view simulation. By predicting this counterintuitive failure mode, the schema saves the practitioner from a whole category of well-intentioned but counterproductive effort, directing attention instead to the one move — importing the outside view — that actually shifts the estimate.

Abstract Reasoning¶

The fallacy permits a structural question: which class of forecasts about novel goal-directed action by an agent is unbiased? The answer is those built from the empirical distribution of comparable past attempts, not those built from the agent's internal simulation of the path to completion. This is a substrate-independent commitment about how forecasts of intentional action should be constructed, whether the agent is human, organizational, or algorithmic. It also yields a structural meta-claim: the bias does not require a defective forecaster, only a reasonable process — simulate the steps, sum the durations — operating on a reference class that structurally omits the unimaginable. And it yields a counterintuitive prediction: the bias should be largest where the inside-view simulation is most detailed and internally coherent, because detail and coherence raise the planner's confidence without addressing the blind spot — so domain experts are particularly vulnerable rather than particularly protected, a prediction borne out in the empirical record. These inferences are about the structure of forecasting strategies, and they apply to any planning agent with a reference class it is not disciplined to use.

Knowledge Transfer¶

The transferable content is an intervention family, all aimed at forcing the outside view into the forecast. Reference-class forecasting explicitly computes the empirical distribution of completion times or costs for a class of comparable past tasks and adjusts the inside-view estimate toward it by an empirically fitted factor. The pre-mortem asks the team, before commitment, to imagine the project has already failed and to write its post-mortem, forcing the inside-view simulation to visit the failure modes it would otherwise skip. Calibration training with explicit feedback yields real but slow and partial gains, since the bias is structural enough to resist individual-level fixes. Independent estimators who are not the executors are less committed to the simulation and so less prone to inside-view substitution — an organizational rather than individual move. Empirical buffers scale the inside-view estimate by the historical ratio of realized-to-estimated for the reference class, of which folk rules ("multiply the estimate by π," "double the cooking time") are rough applications. Optimism-bias uplift embeds these adjustments into regulation, requiring major-project applications to add empirically derived uplift factors. And small-batch planning shortens each plan's horizon so the inside-view simulation has fewer hidden surprises, re-planning frequently rather than committing to a long forecast.

The structural roles map across substrates. The forecasting agent is the student, project office, surgeon, or commander; the two strategies are inside-view simulation and outside-view reference; the default substitution is the pull toward the vivid, self-coherent simulation; the blind spot is the omitted long tail of interruptions; and the structural fix is whatever forces the empirical distribution into the estimate. A megaproject planner applying a regulatory uplift factor, a hospital booking operating-room time from per-surgeon history, and a software lead shortening sprints to limit forecast horizon are performing the same structural act: replacing or correcting an inside-view simulation with an outside-view reference. The diagnostic — which strategy produced this estimate, and what would the reference class predict? — travels unchanged across cognition, organizations, engineering, software, medicine, and military planning. Because the fix is the same everywhere, a practitioner who has tamed the fallacy in one domain — by building a reference class, running pre-mortems, or delegating estimation to a non-committed party — can import the whole repertoire into a domain that knows only that its plans keep running long.

Examples¶

Formal/abstract¶

Consider the canonical thesis-completion study reframed as a contest between two estimators. Students forecasting when they will finish their honors theses are the forecasting agents; the two strategies are inside-view simulation ("I'll write the lit review in two weeks, then the methods in one...") and outside-view reference (the empirical distribution of when comparable students actually finished). Asked for best-case, most-likely, and even worst-case dates, students default to the inside-view simulation and produce estimates clustered well below realized completion times. The structural blind spot is decisive: the simulation can only enumerate steps the student can imagine — it cannot contain the advisor's sabbatical, the corrupted data file, the two weeks lost to flu — because those are precisely the unimaginable interruptions, and the reference class captures them only empirically, as a long right tail. The directional bias is therefore not symmetric error but systematic underestimate, and it shows feedback resistance: a student who agrees past theses ran long, and that hers is comparable, re-simulates and produces another low number. The structural prediction that distinguishes this prime from generic optimism is sharp and counterintuitive: adding inside-view detail — a finer task breakdown, a tighter Gantt chart — worsens the estimate by deepening the simulation without touching its blind spot. The only corrective that moves the number is structural: compute the reference-class distribution and adjust the estimate toward it.

Mapped back: the students are the forecasting agents, plan-simulation versus historical-distribution are the two strategies, the omitted interruptions are the blind spot, and the failure of feedback to fix it is the feedback-resistance signature — the fallacy worked end-to-end.

Applied/industry¶

Three institutional substrates show the same structure with the corrective built into practice. In transport megaprojects, decades of records show a right-skewed pattern of time and cost overruns that does not improve despite repeated promises of better planning — the inside-view promoter's simulation omits the geological surprise, the litigation, the scope change. The intervention is the prime's outside-view fix made regulatory: reference-class forecasting with an empirically fitted optimism-bias uplift, so a major-project application must scale its inside-view estimate by the historical realized-to-estimated ratio for its class before approval. In hospital operating-room scheduling, surgeons' inside-view estimates of procedure duration run systematically low, so booking systems now draw per-surgeon, per-procedure empirical distributions — an outside-view reference replacing the simulation entirely. In software delivery, teams that decompose a release into ever-finer tasks still ship late, because decomposition deepens the inside view; the structural response is small-batch planning — shortening each plan's horizon so fewer unimaginable surprises can accumulate within it — combined with velocity tracking, which is functionally an outside-view reference class built from the team's own history. Across all three, the diagnostic is identical: ask which strategy produced the estimate and what the reference class predicts, then force the empirical distribution into the forecast.

Mapped back: megaproject promoters, surgeons, and software teams are the forecasting agents; promotional simulation, surgeon intuition, and task decomposition are the inside views; and regulatory uplift, per-surgeon history, and velocity tracking are the outside-view references that correct the downward-biased estimate across infrastructure, medicine, and engineering.

Structural Tensions¶

T1 — Reference Class Construction (scopal). The corrective is "use the outside view," but the reference class is itself a choice, and a too-broad class dilutes signal while a too-narrow one ("projects exactly like mine") collapses back into the inside view by way of the same "my case is different" reflex. The frame demands a reference class but does not fix its boundary. Failure mode: gerrymandering the class until it predicts what the inside view already wanted. Diagnostic: was the class defined by structural features fixed before seeing this project, or curated to match a desired estimate?

T2 — Bias versus Strategic Misrepresentation (sign/direction). The prime attributes underestimation to a cognitive blind spot, but in funded projects the same low estimate is often deliberate — a promoter low-balls to win approval, knowing overrun is recoverable. The competing prime is strategic misrepresentation / agency problem. Failure mode: prescribing de-biasing interventions (pre-mortems, reference classes) against what is actually an incentive problem, which they cannot fix because the forecaster is not mistaken. Diagnostic: does the forecaster bear the cost of the overrun? If not, suspect motive, not bias.

T3 — Padding Overshoot (measurement). Forcing an outside-view uplift can overshoot: a fixed optimism multiplier applied indiscriminately inflates estimates for the genuinely-novel-and-fast project, and padded estimates then invite Parkinsonian fill. The correction has no built-in stop. Failure mode: a blanket "double every estimate" rule that makes the organization uncompetitive and trains everyone to pad pre-emptively. Diagnostic: is the uplift calibrated to the realized distribution of this class, or a one-size reflex? And does padded time get consumed regardless (the Parkinson coupling)?

T4 — Novelty Boundary (scopal). The fallacy is specified for novel tasks where no good reference class exists in the planner's experience — but for genuinely unprecedented work, the outside view is unavailable by construction, and the inside view, flawed as it is, may be all there is. The prime's own scope condition (novelty) undercuts its own remedy (reference class). Failure mode: demanding a reference class for a true first-of-its-kind effort and either fabricating a false one or paralyzing the estimate. Diagnostic: does a comparable-attempt population actually exist? If not, the corrective is decomposition-plus-buffer, not reference-class forecasting.

T5 — Detail Worsens the Estimate (temporal). The counterintuitive structural claim is that finer inside-view decomposition deepens the bias rather than reducing it — yet decomposition is the universal instinct when a plan looks shaky, and is genuinely useful for execution even as it harms forecasting. The same activity helps one goal and hurts the other. Failure mode: responding to a suspect estimate by breaking it down further, producing a more confident and more wrong number. Diagnostic: separate planning-for-doing from estimating-the-total; never let the task breakdown become the duration forecast.

T6 — Forecast Resistance to Its Own Correction (coupling). Because the bias is feedback-resistant, even an agent who knows the planning fallacy will re-simulate and produce another low number — the correction must be structural (delegate estimation to a non-committed party, mandate the uplift) rather than attitudinal. But delegating estimation severs it from the knowledge held only by the committed planner. Failure mode: an outside estimator who avoids the optimism bias but lacks the project-specific information that would have improved accuracy, trading one error for another. Diagnostic: is the estimation authority separated from execution commitment without losing access to the planner's private information?

Structural–Framed Character¶

The planning fallacy sits near the framed pole of the structural–framed spectrum, consistent with its aggregate of 0.9 — among the most framed primes in the catalog. There is a thin relational skeleton in the abstract — two forecasting strategies, inside-view simulation and outside-view reference class, producing systematically different estimates with the agent defaulting to the biased one — but nearly every diagnostic loads toward framed, and the pattern cannot be lifted out of its human-cognitive home.

The home vocabulary travels heavily: "inside view," "outside view," "reference class," "pre-mortem," "optimism-bias uplift" are not generic terms one swaps for substrate-local words — they are the constitutive lexicon of a specific judgment-and-decision-making research program. Institutional origin is explicit: the planning fallacy is a named cognitive bias introduced by Kahneman and Tversky, and it remains tied to that literature and to professional forecasting practice. The prime is thoroughly human-practice-bound — it presupposes a planning agent that simulates a path, imagines steps, and holds beliefs about its own task; megaprojects, software releases, theses, surgery, and military operations are all human-practice substrates, and there is no non-agentic physical or biological instantiation, since nothing without intentional forecasting can commit the fallacy. And invoking it imports a whole interpretive frame about how forecasts ought to be constructed — from the empirical distribution, not the simulated path — rather than recognizing a pattern wired into an indifferent system.

The one diagnostic at the midpoint is evaluative weight: "fallacy" carries a charge of error, and the prime names a thing to be corrected, yet the entry is careful that the bias arises from an otherwise-reasonable process operating on a structurally unrepresentative reference class, not from carelessness — so the disapproval is muted rather than absolute. The genuine relational core (two strategies, one biased) is real, but it is inseparable from the cognitive-forecasting frame that gives it content, which is exactly why the prime sits where the grade places it.

Substrate Independence¶

The planning fallacy is a low-to-moderately substrate-independent prime — composite 2 / 5 on the substrate-independence scale. The pattern — forecasters of a novel task defaulting to inside-view simulation over the outside-view reference class and so systematically underestimating time, cost, and risk — does show up across several arenas, giving a domain breadth of 3: infrastructure megaprojects, software development schedules, surgical and clinical time estimates, and military operational planning. But that breadth is illusory in the substrate sense, because every one of those arenas is a human-practice forecasting setting: the fallacy presupposes a cognitive agent constructing a plan by mental simulation, with no physical, biological, or non-cognitive instantiation, which holds structural abstraction down to 2. Transfer evidence is a 3 — the reference-class-forecasting correction is well documented and the underestimation is measured across megaprojects and software, but the "transfer" is the same cognitive bias re-observed in different planning domains rather than a medium-crossing formalism. The composite is a 2: a robustly attested but firmly human-practice-bound phenomenon.

Composite substrate independence — 2 / 5
Domain breadth — 3 / 5
Structural abstraction — 2 / 5
Transfer evidence — 3 / 5

Relationships to Other Primes¶

Parents (2) — more general patterns this builds on

Planning Fallacy is a kind of, typical Bias

A named systematic directional forecasting error; a bias. Secondary parent (optimism_bias is itself a child of bias).
Planning Fallacy is a kind of, typical Optimism Bias

The file: planning_fallacy is 'optimism bias as it manifests in forecasts of one's own goal-directed tasks' — the directional skew specialized to a specific MECHANISM (inside-view simulation substituting for the outside-view reference class). A specialization of optimism_bias with extra mechanistic content (feedback-resistance, detail-worsens-it).

Path to root: Planning Fallacy → Bias

Neighborhood in Abstraction Space¶

Planning Fallacy sits among the more crowded primes in the catalog (37^th percentile for distinctiveness): several abstractions describe nearly the same structure, so a description that fits it will tend to fit its neighbors too — transporting it usually means disambiguating within this family rather than landing on it exactly.

Family — Anticipation & Forward Models (15 primes)

Nearest neighbors

Computed from structural-signature embeddings · 2026-06-14

Not to Be Confused With¶

The planning fallacy is most often collapsed into optimism_bias, and the two are genuinely close — the planning fallacy is, in part, optimism bias as it manifests in forecasts of one's own goal-directed tasks. The distinction worth preserving is that optimism bias names a direction (a general skew toward favorable expectations across health, finance, relationships, and risk) while the planning fallacy names a mechanism (inside-view simulation substituting for the outside-view reference class). The mechanistic specification carries content that the directional label does not. It predicts that the bias survives feedback and expertise, because the cause is the simulation strategy's structural blind spot rather than a mood or disposition that information could correct. It predicts that adding inside-view detail worsens the estimate — a prediction optimism bias alone would never generate, since more careful thought ought to reduce a mere disposition. And it specifies the corrective precisely: not "be more realistic" but "import the empirical distribution of comparable attempts." A practitioner who diagnoses generic optimism bias will reach for de-biasing exhortation or calibration pep-talks; one who diagnoses the planning fallacy reaches for reference-class forecasting and pre-mortems, because the problem is a wrong forecasting strategy, not a wrong attitude.

A second, sharper confusion is with the agency_problem (and its cousin, strategic misrepresentation), because the two produce the same observable — a forecast that comes in systematically below realized cost — yet have opposite causes and opposite fixes. In the planning fallacy the forecaster is sincerely mistaken: the inside-view simulation genuinely omits the long tail, and the planner believes the low number. In the agency-problem case the forecaster is not mistaken at all; a promoter who must win funding deliberately low-balls the estimate, knowing the overrun will be absorbed later and borne by someone else. The discriminator is incentive and cost-bearing: does the forecaster personally bear the cost of the overrun? If yes, suspect the fallacy (a sincere blind spot); if the cost falls on a principal while the agent captures the approval, suspect strategic misrepresentation. The fix diverges accordingly. De-biasing interventions — pre-mortems, reference classes, calibration — cannot touch an agency problem, because the forecaster is not in error; the appropriate response there is incentive realignment, skin-in-the-game, or independent review. Conversely, contractual incentive fixes do nothing for a sincere cognitive blind spot. Confusing the two means prescribing a cognitive remedy for an incentive disease, or vice versa — a common and expensive misdiagnosis in megaproject governance.

Finally the planning fallacy must be separated from scenario_planning, its embedding-nearest neighbor, with which it is confused only because they co-occur — but they sit on opposite sides of the problem. Scenario planning is part of the remedy family: the disciplined construction of multiple alternative futures, including adverse ones, precisely to drag the forecaster's attention off a single self-coherent path. The planning fallacy is the condition that makes scenario planning necessary — the default collapse onto one inside-view simulation that omits the unimagined. To treat scenario planning as an instance of the fallacy, or the fallacy as a kind of scenario planning, inverts the relationship between disease and cure. The pre-mortem, branch-and-sequel analysis, and reference-class forecasting are all scenario-planning-adjacent correctives that work by forcing the outside view or the failure modes into a forecast that the fallacy would otherwise leave smooth and singular.

These distinctions matter because each names a different intervention. Confusing the fallacy with optimism bias leads to attitudinal de-biasing that the feedback-resistant structure defeats; confusing it with an agency problem leads to applying cognitive remedies to an incentive disease (or the reverse); and confusing it with scenario planning inverts the relationship between the failure and its cure. The fallacy's signature contribution — forecast novel action from the empirical distribution of comparable attempts, never from the simulated path — is the one move none of these neighbors supplies on its own.

Solution Archetypes¶

No catalogued solution archetypes reference this prime yet.