Why it matters

Most forecasting quietly assumes the future is a fixed thing you cannot yet see — a number behind a curtain, and the job is to estimate it well. For a great many questions that is exactly right: next quarter’s demand, the failure rate of a part, the path of a hurricane over the next three days. But some futures are not like that. They are not merely uncertain; they are deeply uncertain — you cannot even agree on the right model or the right probabilities — and they are reflexive: the forecast and the actions it triggers reach back and change the future being forecast. For those questions, projecting a trend or putting a number behind the curtain is not just imprecise. It quietly misrepresents the kind of thing you are looking at. Wicked Futures is the mode for that kind of future — the one that mutates as you study and act on it.

For example: forecast a pandemic’s death toll and publish it. If the number is frightening, people stay home, governments act, and the toll comes in far lower than you said — so were you wrong? If the number is reassuring, people relax, and the toll comes in higher. The act of forecasting changed the thing forecast. There is no untouched ground truth to have been right about. A point forecast hides exactly the feature that matters most; what you actually need is a map of which futures are reachable, which forces bend the path between them, and which early signals tell you which way the world is tipping.

  • What it reveals. The structure of a deeply uncertain, reflexive future — the handful of internally coherent futures that are genuinely reachable, the deep uncertainties and feedback effects that route between them, and the contested framings underneath, rather than a single projected line.
  • How it changes the read. You stop asking “what is the most likely outcome?” and start asking “which futures are reachable, what forces move us between them, and what would I have to watch to know which way the world is tipping?”
  • When to foreground it. A forward-looking question where the uncertainty is deep (you cannot honestly settle on one model or one probability), the system is reflexive (forecasting and acting change the outcome), and the framing itself is contested — climate, AI governance, pandemic response, the long arc of an institution.
  • What you’d miss without it. That the future in question has no stable target to forecast toward; treat it like a hurricane track and you get a confident line through a space that is actually branching, contested, and bent by your own response.
  • Where it misleads. Forced onto a future that is not wicked — a bounded, well-modelled question with stable structure — it manufactures false drama and wide branching where an ordinary forecast or a clean set of scenarios would have served better and been more honest.

Realtime examples

See real, dated analyses where this mode mapped a deeply uncertain future behind a story in the news → Wicked Futures on Main Street Independent

How to invoke it in Ora

You have a forward-looking question where the uncertainty runs deep — you cannot honestly pick one model or one probability — the system feeds back on itself, and even the framing is contested. You want the shape of that future mapped: the reachable futures, the forces that route between them, and the signals that tell you which way it is tipping — not a single projected line.

Describe the question and its horizon, and ask:

“Wicked future on [domain] over [years]: long-horizon scenarios with probabilities, plus the deep uncertainties and feedback effects that route between them, and pre-mortem failure pathways for the leading scenarios.”

The phrase wicked future, together with a named domain, a time horizon, and a request for scenarios with probabilities plus failure pathways, is what routes you here. Bring three things: the forward question stated plainly, the time horizon (a 2026–2031 window forecasts very differently from a 2026–2046 one), and any critical uncertainties or interventions you already have in mind. The richer the question — “the four-year residential college over 2026–2040, with AI tutoring at scale, the demographic cliff arriving, and credential-signal erosion converging” beats “the future of college” — the sharper the map.

One boundary worth knowing. If your question is a present-tense one — a tangle of stakeholders, interests, and power with no agreed problem definition, here and now — that is the Wicked Problems mode, which structures a contested present. Wicked Futures takes the same wicked-problem dynamics and points them forward in time, at a future that is deeply uncertain and reflexive. Same family of difficulty; one faces the present, the other the future.

How it works

Start with the kind of future this mode is not for, because the contrast is the whole point. A meteorologist forecasting a hurricane three days out is working under ordinary uncertainty. The storm has a real track. The forecaster does not know it exactly, so they run many models and report a cone of probability — but the cone narrows as the storm approaches, the storm does not care what the forecast says, and when it makes landfall everyone can check who was right. There is a ground truth, and the forecast converges toward it. That is uncertainty you can put a number on, and an ordinary forecast handles it well.

Now change the question to something like pandemic response, or how a society governs a powerful new technology, or the long path of the climate. Three things break at once. First, the uncertainty stops being a number. It is not that you know the model and are unsure of the inputs; you are unsure of the model itself — which forces dominate, how they interact, whether the rules that held last decade still hold. That is what the deep-uncertainty literature calls deep uncertainty: not “we don’t know the value” but “we can’t even agree on the structure or the probabilities.” Second, the system becomes reflexive — it reacts to being forecast. Publish a frightening pandemic number and people change their behaviour, which changes the number, so the forecast partly unmakes itself. The hurricane ignores the forecast; the society does not. Third, the framing itself is contested — reasonable people disagree not just about what will happen but about what the question even is, what counts as a good outcome, where the boundary of the problem sits. Whose climate future, measured how, traded off against what?

These three features — deep uncertainty, reflexivity, contested framing — are the marks Horst Rittel and Melvin Webber identified in 1973 when they named wicked problems: problems with no definitive formulation, no stopping rule, no true-or-false answer, where every attempt at a solution changes the problem. Rittel and Webber were describing tangled problems in the present. Wicked Futures takes that same diagnosis and turns it toward the future. A future can be wicked in exactly their sense: there is no settled formulation of the question, no clean target to forecast toward, and the act of forecasting and acting reshapes what unfolds.

So what do you do when there is no number behind the curtain? You stop trying to predict the future and start trying to map it. Instead of one projected line, you lay out the small set of futures that are genuinely reachable — three to five internally coherent stories, each a different way the world could plausibly cohere — and you make them genuinely different by building them along the axes you are most deeply uncertain about. (Will the technology mostly augment people, or replace whole workflows? Will institutions react in eighteen months, or never?) Then you do the thing that distinguishes this from ordinary storytelling: you trace the divergence points — the specific forks where the world commits to one branch over another — and for each you name a leading indicator, an early signal you could actually watch. And because the future is reflexive, you run a pre-mortem against the futures you would most likely plan for: you stand at the far end, assume the plan failed, and ask what feedback loop or blind spot did it in. The deliverable is not a prediction. It is a map of a contested, branching, self-bending future, with the signposts marked.

The discipline that makes the map honest is a refusal: no single most-likely future. Even when the probabilities cluster, the mode will not crown a winner and let you stop watching the others, because the whole reason this future is wicked is that it can still tip — and your own choices are one of the things that tips it. The point of the map is to keep several futures live and tell you what to watch, so that which future you are heading into becomes visible as the signals come in, while there is still time to act. A confident point forecast would feel more decisive and would quietly throw that away.

Framework & implementation

This section uses Ora’s own terms for the parts of an analysis, so that if you open the actual mode file they line up. Each is glossed in plain language on first use.

Pipeline execution

Wicked Futures is the heaviest forward-looking mode in the future-exploration territory — a molecular mode, meaning it is composed from several smaller analyses rather than being a single pass. It runs at Gear 4, Ora’s most thorough setting: a Depth analyst and a Breadth analyst work the question in parallel and then critique each other (cross-adversarial evaluation) before a consolidator integrates the result.

The mode runs three component analyses and then three synthesis stages that braid them together. The components are scenario-planning (the internally coherent narrative futures), probabilistic-forecasting (calibrated probability bands on those futures), and pre-mortem-action (standing at the far end of a future and asking how it failed — the adversarial-future stress test). The synthesis stages then (1) overlay probabilities onto the scenarios, (2) run the pre-mortem stress test against the leading scenarios — the ones you are most likely to plan for, not just against a candidate intervention — and (3) fold everything into a single integrated forward architecture that says something no one component could say alone: which futures share an early signal, where the divergence points are, and what to monitor.

Two disciplines are load-bearing in the pipeline. The integration requirement (Ora’s term is the silo-aggregation guard) forbids the mode from handing back three side-by-side reports — scenarios here, probabilities there, failure pathways over there; the synthesis must produce claims that only the combination yields. And the no-most-likely discipline preserves scenario planning’s anti-prediction stance: even when probability bands cluster, the mode refuses to designate one future as most likely, forcing you to monitor the divergence points rather than commit early. A fourth possible component — backcasting, working backward from a desired future to the interventions it would require — is deliberately deferred, and the mode flags that gap visibly in the output rather than quietly substituting a weaker stand-in.

Output contract

The deliverable is a fixed set of sections, so the map is auditable rather than a loose narrative: Forward Question and Horizon (the question stated plainly, with its time window), Scenario Set with Probability Bands (three to five internally coherent futures, each with a narrative, a type, and a probability range that is allowed to be wide), Divergence Points (the specific forks that route between scenarios, each with a leading indicator you could watch), Failure Pathway Stress Test Findings (a pre-mortem run against the leading scenarios — for each, the failure assumed already to have happened, its leading scenario, its early indicators, and whether it is recoverable), Integrated Forward Architecture (the synthesis: which futures share a signal, what the most probable branches have in common, what to watch), a Constructive-Future Gap-Flag (the explicit note that backcasting was deferred), and Divergence-Points-to-Monitor (the consolidated watch-list of signals and the patterns that route each toward a particular future).

Origin and evidence

The mode marries a diagnosis to a method. The diagnosis is Rittel and Webber’s Dilemmas in a General Theory of Planning (1973), which named the wicked problem — no definitive formulation, no stopping rule, no true-or-false test, every intervention reshaping the problem — and is the source of the claim that some questions cannot be forecast toward a fixed target. Wicked Futures applies that diagnosis forward in time.

The method comes from two literatures. The first is decision making under deep uncertainty, the strand of policy analysis that asks how to act when you cannot agree on a model or assign probabilities. Its founding modern statement is Lempert, Popper and Bankes’ Shaping the Next One Hundred Years (2003), which introduced robust decision making — seeking strategies that perform acceptably across many futures rather than betting on the one you predict — and the field’s later synthesis, Decision Making under Deep Uncertainty (Marchau, Walker, Bloemen and Popper, eds., 2019), which collects robust and adaptive methods for exactly the situations where ordinary forecasting fails. The second is the scenarios tradition’s reckoning with these questions: Ramírez and Selin’s Plausibility and Probability in Scenario Planning (2014) argues that for genuinely uncertain, contested futures the right currency is plausibility — internally coherent, reachable futures — rather than probability alone, which is the intellectual backbone of the mode’s no-most-likely discipline. Wicked Futures is the assembly of these: Rittel and Webber’s wicked-problem structure, the deep-uncertainty stance, and scenario plausibility, combined into one forward map.

Applications and common uses

  • Technology governance. How a society governs a powerful, fast-moving technology — AI, biotech — where the technology, the rules, and public reaction all bend each other.
  • Long-horizon socio-economic forecasts. Labor displacement, demographic shifts, the future of an institution (the four-year college, a profession), where multiple deep uncertainties converge over a decade or more.
  • Climate and environment. Futures where physical, political, and economic systems interact, the framing is contested, and policy choices reshape the trajectory.
  • Pandemic and public-health response. Reflexive futures par excellence — the forecast changes behaviour, which changes the outcome — where you need reachable futures and signals, not a single curve.
  • Geopolitical realignment. Multi-actor futures (a great-power rivalry, a regional order) where the framing itself is contested and the system reacts to being analysed.

Failure modes and when not to use it

  • Trend extrapolation creeping back in. The deepest pull is to quietly project the present forward and call it a scenario. The mode guards against it by forcing the futures apart along explicit deep-uncertainty axes — if the scenarios differ only in degree, the axes were not deep enough.
  • Silo aggregation. Returning scenarios, probabilities, and failure pathways as three side-by-side reports that never speak to each other. The integration requirement is the guard; the test is whether the synthesis says something no single component could.
  • False precision on the probabilities. Reporting tight probability bands on a genuinely deep-uncertainty question is its own failure. Wide bands (5%–40%) here are not sloppiness — they are an honest picture of the structure, and the mode is built to widen, not narrow, the further out the horizon runs.
  • Suppressing the gap-flag. Quietly substituting a weak backward-from-a-goal analysis for the deferred backcasting component. The mode surfaces the gap instead.

When not to reach for it. When the difficulty is a present-tense tangle of stakeholders, interests, and power with no agreed problem definition — that is a contested present, not a contested future, and the Wicked Problems mode is the right tool. When the uncertainty is bounded and well-structured — you trust the model and are merely unsure of the inputs — a clean set of futures from scenario-planning, or a calibrated estimate from probabilistic-forecasting, will be sharper and more honest than the full wicked apparatus. And when the future in question simply is a fixed target you can forecast toward (next quarter’s demand, a hurricane track), the branching, contested, self-bending machinery of this mode produces drama where a straight forecast would do.

  • Wicked Problems — the present-tense sibling: the same wicked-problem dynamics (no agreed definition, contested framing, every move reshaping the problem) applied to a tangled present of stakeholders and power, rather than a deeply uncertain future.
  • Scenario Planning — the component for internally coherent narrative futures without the probability overlay or the pre-mortem stress test — the right tool on its own when the uncertainty is real but bounded.
  • Probabilistic Forecasting — the component for calibrated probability bands without the scenario narratives — the right tool when the structure is stable enough to honestly put numbers on.
  • Pre-Mortem Action — the lens this mode loads to stress-test its leading futures: stand at the far end, assume the plan failed, and trace the feedback loop or blind spot that did it in.