Belief being tested: A sufficiently detailed forecast makes every important uncertainty calculable.
A forecast can become more elaborate without becoming more dependable. Additional categories, formulas, and scenarios can describe an assumption more precisely while leaving the assumption unsupported.
The distinction matters because a decision can look measured when the underlying uncertainty remains unresolved.
Frank Knight’s Risk, Uncertainty, and Profit distinguishes measurable risk from uncertainty that lacks a dependable basis for assigning probabilities. Knight also recognizes that practical situations do not always fall neatly into separate classes.
The distinction asks what supports a probability judgment. It does not imply that every unfamiliar event is beyond analysis.
In this invented situation, an owner is considering a service the company has not previously offered. The forecast includes expected customer demand, staffing needs, and the pace at which the work will become routine.
Some assumptions draw on relevant experience. Others depend on customer behavior the company has not observed.
Putting those assumptions into the same spreadsheet does not give the assumptions the same evidentiary standing.
An owner can ask where each expectation came from. Does the company have comparable work? Are the customers comparable? Do the observations include the conditions the new service will face? Is the forecast borrowing a pattern from a situation that differs in an important way?
These are questions about the foundation of the forecast.
For an owner, a useful forecast exposes the decisions that depend on judgment. A forecast becomes less useful when its presentation hides those decisions.
In the invented example, the owner might have sound evidence about the time required for a familiar administrative task. The owner might have little evidence about whether customers will accept a new service arrangement.
The familiar task can support a planning estimate. The customer assumption needs a different kind of attention.
That attention could include asking prospective customers about their actual choices, testing a limited version of the service, or examining why a comparable offering failed. Each action should address a specific uncertainty.
Collecting more information is not automatically progress. Information helps when the information can change a decision or clarify a consequence.
The owner also needs to distinguish an assumption from a commitment. The forecast may assume that demand arrives gradually. A staffing commitment may be difficult to reverse if demand develops differently.
The consequence of being wrong therefore matters alongside the plausibility of the forecast.
A useful review asks what happens if an important assumption fails. What work can stop? What obligation continues? What signal would justify changing course? Who must notice that signal?
Those questions do not convert uncertainty into a known probability. The questions help the company act responsibly despite incomplete knowledge.
Knight’s theory supplies the distinction between kinds of uncertainty. The proposed decision routine is an operating judgment. The routine has to earn its place through the quality of the decisions it supports.
There is no reason to treat uncertainty as permission for intuition without examination. An owner can lack a dependable probability estimate and still compare alternatives, seek contrary evidence, and limit avoidable commitments.
There is also no reason to punish a forecast merely because the actual result differs. A responsible forecast can be wrong. A careless forecast can be lucky. The review should examine what was reasonably known and how the decision used that knowledge.