[!todo] Seed note. A starting point, not a finished note yet.
The Cone of Uncertainty describes how far a software estimate can be off at each phase of a project. Barry Boehm measured the effect at TRW in the early 1980s and Steve McConnell named it. At initial concept, estimates are accurate only to within roughly 4x in either direction, a 16x spread between the plausible low and high, narrowing to about 1.6x after requirements are fixed and converging on the actual near completion. The load-bearing caveat is that the cone narrows only as you actively resolve the decisions that create the variance, not as the calendar advances. A project that has been running for months without pinning down scope sits in the same fat part of the cone it started in, which makes the “cone” a cloud. Seeded from Software Estimation and Forecasting.