The Futures Cone

Most analytical work assumes the future is singular — one forecast, one target, one plan. We build models that output a number, a confidence interval at best, and call it done.¹

The Futures Cone offers a different architecture: a structured way to think about multiple possible outcomes simultaneously. Not as a hedge, but as a tool for decision-making under genuine uncertainty.

The architecture

The cone expands from a single point — now — into four nested regions:

• Possible — everything that could happen, constrained only by physics
• Plausible — what might happen given current knowledge
• Probable — what likely happens if current trends continue
• Preferred — what we want to happen

The interesting zone is where Probable and Preferred overlap. That’s the sweet spot — futures that are both likely and desirable. The gap between them is where strategic work lives.

Use the slider to explore what happens when what’s likely diverges from what’s desirable — and when they converge.²

Why this matters for Data Science

The interesting question for Data Science isn’t prediction accuracy — it’s decision architecture. Which futures are we building toward, and which are we building against?

When we collapse futures into a single prediction, we lose the ability to reason about the shape of possibility itself. The cone doesn’t tell you what will happen. It tells you what kind of thinking you need to do.

I’m not fully convinced “preferred” is the right framing — it implies individual preference where collective intention might be more useful. The cone works best when it’s a shared artifact, not a personal forecast.³

The real power is in the conversation the cone generates. When you put possible, plausible, probable, and preferred side by side, you stop arguing about what will happen and start asking: what kind of future are we actually building?

That’s a fundamentally different question — and most analytical pipelines aren’t built for it.