Futures Cones: Probable, Plausible and Possible Futures Explained

A futures cone is a diagram of how the space of possible futures widens as you move away from the present. A single point — now — opens into a cone divided into bands: probable futures that follow current trajectories, plausible futures that are imaginable but less likely, and possible futures that are unlikely but not ruled out by what we understand. Outside the cone is what we would call impossible. It is one of the most reproduced images in foresight, and one of the most misread.

What the cone actually represents

The cone makes one argument, and it is about time rather than probability. There is a single present, so the diagram begins at a point. Move right and the number of states the world could be in grows, so it opens. Everything else — bands, labels, shading — is annotation on that.

One consequence matters most, because it is where the misreadings begin: the cone has no vertical axis. Nothing is measured. Distance from the centre line means only "further from the extrapolated baseline". It is not a probability distribution, though it is read as one constantly.

The bands: probable, plausible, possible, impossible

The bands are not evenly occupied. Almost all professional planning sits in the probable, which is narrow because continuity is a narrow assumption. The plausible needs only that some current arrangement — a regulation, a price, a habit — changes. The possible needs something we do not have yet, and is where most speculative design aims.

Impossible is the band that ages worst. The history of the field is largely a history of things migrating out of it and into probable inside a working lifetime. Joseph Voros, following Stuart Candy, added a further band called preposterous for the futures a room dismisses with a laugh — an honest acknowledgement that "impossible" hides a lot of assumptions.

Where the diagram came from

• •Charles Taylor, US Army War College, late 1980s — a "cone of plausibility" for military planning.
• •Trevor Hancock and Clement Bezold, "Possible Futures, Preferable Futures" (Healthcare Forum Journal, 1994) — the version closest to today's, with preferability laid over the nested bands.
• •Roy Amara, at the Institute for the Future in the early 1980s, framed the futures field as concerned with the possible, the probable and the preferable. He is better remembered now for Amara's Law, which matters here too.
• •Wendell Bell, Foundations of Futures Studies (1997) — that trio's most careful philosophical treatment.
• •Joseph Voros, from his 2003 paper in Foresight onwards, extended the diagram and traced its history honestly rather than claiming it. He is why most practitioners have seen it.

Dunne and Raby carried it into design education as the "PPPP" diagram in Speculative Everything.

Preferable is not a probability

This is the distinction most people get wrong, and getting it wrong quietly wrecks strategy work.

Probable, plausible and possible are claims about knowledge: they answer what does our understanding permit? Preferable answers a different question — what do we want? That is a values claim. No amount of evidence settles it, and no amount of desire moves the other three.

Which is why Voros does not draw preferability as a fourth nested band, but as a separate shape lying across the others: it can overlap the probable, spill through the plausible, sit mostly in the possible, or fall outside the cone altogether. The geometry is deliberate.

Two errors follow from forgetting that.

Treating preferable as the good end of probable. A team names, as its preferred future, the best realistic outcome of the current trajectory — collapsing a values question into a feasibility question. The result is not strategy but optimism about the status quo. A reliable test: if your preferable future always sits inside your probable band, you are describing what you expect and calling it a preference. A preferable future outside the possible band is not a mistake but a direction, and naming it is often the only way to discover what would have to change.

Leaving out "preferable to whom". Preference has an owner. A future preferable to a logistics operator is not automatically preferable to its drivers, or to the towns its warehouses sit beside. Any real organisation contains several preferable futures in open conflict, and the diagram's single shaded region conceals them.

One illustration makes this concrete. A probable future in which nearly every car is electric is not the same as a preferable future in which far fewer people need a car at all. The difference is not likelihood but a different idea of what a good settlement between people and movement looks like. Conflating the two is how a decarbonisation strategy becomes a procurement plan.

Why we draw it on an exponential curve

Straight sides encode an assumption: that uncertainty accumulates at a constant rate. We do not think it does, so in our teaching the boundaries are exponential curves. Change compounds, and Amara's Law names the consequence — we overestimate the effect of a technology in the short run and underestimate it in the long run. A straight-sided cone is a picture of exactly that linear estimate.

Three things change when you redraw it.

The near term is tighter than the straight cone suggests. Within two or three years, probable and plausible are nearly the same shape. An organisation planning on a two-year cycle is not being unimaginative; at that horizon there is little room. The problem is treating it as the only horizon.

The bands separate late, and then fast. For most sectors, somewhere past ten years, the gap between probable and possible stops being a matter of degree and becomes one of kind. That is why speculative work needs a long horizon — not because the far future is more interesting, but because it is the first place the bands are far enough apart to have anywhere to go.

The probable band stops being an evidence claim. On a straight cone, "probable at five years" and "probable at thirty years" are the same statement with a wider error bar. On an exponential curve they are not: the further right you look, the more the probable reflects the limits of your imagination rather than your evidence.

Solar photovoltaics is the cleanest illustration. Through the 2000s and 2010s the central projections in mainstream energy outlooks — the documents that defined the probable band for an entire industry — were revised upward edition after edition, because deployment kept overshooting them. What happened would, from a 2005 vantage, have been drawn near the outer edge of possible, and required no new physics. Those forecasts were not wrong about the technology. They were wrong about the shape of the cone.

What the cone cannot do

It is a thinking aid, not a measurement. No probabilities, no confidence intervals, no way to validate where anything has been placed. Attaching percentages to the bands misunderstands the object.

The boundaries are not real. They are drawn by the people in the room, out of what those people happen to know. A future that is plausible to an epidemiologist is fantasy to a retail banker. Change the room and the bands move. The cone pictures a group's imagination, not the world.

It invites false confidence about the probable. Drawing continuity as a narrow band down the centre makes it look like the safe bet. But "the world continues broadly as it is" is a demanding claim: it requires every current arrangement to hold at once. Any single alternative may be less likely than continuity; continuity is not more likely than all of them together — see NKD_6 — Against Forecast.

Using it well

The cone's job is argumentative rather than analytical. It does not generate futures; other tools do that. It makes explicit where a piece of work is aiming, early enough that people can disagree cheaply.

Arguing about where a project should aim. Before a futures project starts, have everyone say which band the output should land in. Teams that skip this find the disagreement at the final presentation, when the strategy director wanted probable and the design team delivered possible.

Noticing when a team only ever works inside the probable band. Ask people to place their live projects on the cone. In most organisations — including plenty with "innovation" or "futures" in the team name — the cluster is tight and hugs the centre line. That is not a moral failing; the probable is where revenue lives. But a portfolio with nothing outside it cannot see a discontinuity coming.

Where to take this next

The cone does the framing everything else depends on: it separates what you expect from what you can imagine from what you want. It does not tell you how to build the future in question, or what to do with it.

If you want the rest — including the reasoning behind the exponential curve, and what changes in a project once you adopt it — Speculative Design Basics covers it at your own pace.

Key takeaways

• •The cone pictures uncertainty accumulating over time: one present, many futures. It has no vertical axis and measures nothing.
• •Probable, plausible and possible are claims about knowledge. Preferable is a claim about values, on a different axis — it can sit anywhere, or outside the cone.
• •The commonest error is treating "preferable" as the good end of "probable", turning a values question into a feasibility question.
• •The lineage runs through Taylor, Hancock and Bezold, Amara and Bell, popularised by Joseph Voros. Nobody owns it.
• •We draw the bands as exponential curves, because possibility does not open at a constant rate — which is why speculative work needs a horizon beyond ten years.
• •Its best use is diagnostic: it shows whether a team only ever works inside the probable.