What is a forecast cone?

A forecast cone is a picture of uncertainty. Rather than claiming a market will reach one specific number, it shows the range of prices that market could plausibly reach, and how that range widens the further ahead you look. It is the shape uncertainty actually has.

Why the cone widens

Over five minutes, a market has time to move very little. Over five years, it has time to move enormously. Uncertainty grows with the square root of time — quadruple the horizon and the band roughly doubles in width. That is why the cone flares out to the right rather than running parallel: it is a direct picture of how much less you can know about next year than about the next hour.

How Coneview builds one

• •We measure how volatile the instrument has actually been, using its recent price history.
• •We then simulate 20,000 possible futures for it — a Monte Carlo simulation, where each run is one plausible path the price could take.
• •We sort the outcomes and read off the band containing the middle 80% of them: the 10th to 90th percentile. That shaded region is the cone.

So when the cone says 3,970 to 4,308 for gold over a week, it means: in roughly four of every five simulated futures, the price finished somewhere inside that band. In one of five, it finished outside it — which is not a failure of the model but a stated, expected part of it.

The dashed line in the middle

The dashed line is the median — the 50th percentile, the outcome with as many simulated futures above it as below. It is emphatically not a prediction. It is the middle of a distribution, and the distribution is the actual answer. If someone quotes you only the median and hides the band, they have thrown away everything that mattered.

Why not just give a price target?

Because a single number is a claim to knowledge nobody has. "Gold will hit 4,300" is unfalsifiable theatre: if it does, the forecaster looks prescient; if it doesn't, the date simply moves. A range with a stated probability is a claim that can be checked— over enough forecasts, prices should land inside the 80% band about 80% of the time. If they don't, the model is wrong, and you can prove it.

That is the trade Coneview makes. We give up the false comfort of a precise number in exchange for a forecast you can actually hold us to. Our public accuracy scorecard is where we do exactly that.

Reading a cone in practice

• •A wide cone means genuine uncertainty — the instrument has been volatile, and the honest answer is that a broad range of outcomes is plausible.
• •A narrow cone means the recent range has been calm. It does not mean a large move is impossible, only that it has been rarer lately.
• •The cone is anchored to now. Its edges are fixed at the moment the forecast was made. As live price moves within the band, the distance to each edge changes — the edges themselves do not chase it.

Next: how to read the P10 and P90 edges, and what makes a probability trustworthy.