Conviction without certainty
Serious research does not eliminate uncertainty. It makes the basis of a decision more visible and more open to revision.
Perspectives · 2026-11-07
A model is valuable because it leaves things out. By simplifying reality, it allows a relationship to be examined. The same simplification is also a limitation: an answer can only be understood in relation to the assumptions and structure that produced it.
What a model can clarify
NIST's work on measurement uncertainty describes probability and statistical tools as foundations for expressing and interpreting uncertainty within measurement science. It discusses the role of models in relating quantities and observations. [5] Measurement science is not investment forecasting, but it offers a useful discipline: an estimate should not be separated from an account of how it was formed.
In an investment context, that lesson is best treated as an analogy rather than an imported guarantee of rigor. A forecast may combine data, judgments and assumptions that cannot all be measured in the same way. Giving the output more decimal places does not resolve those differences.
The obligation to distinguish
Useful research distinguishes what has been observed from what has been inferred. It also distinguishes an explanation from a story that merely fits the outcome after the fact. These distinctions make it possible to ask where confidence comes from and what evidence would change it.
One practical question is especially revealing: which assumption is doing the most work? If a conclusion depends heavily on a favorable development, that dependence belongs at the center of the discussion. It should not disappear behind the sophistication of the presentation.
Scenarios as intellectual discipline
A scenario is a structured question about conditions. What changes if a project takes longer? If a customer adopts a substitute? If a crucial input becomes less available? The purpose is not to decorate a central forecast with alternatives. It is to understand how different developments could alter the logic of a decision.
This gives patience a more demanding meaning. Patience is not a refusal to act, and it is not a refusal to reconsider. It is the ability to let the quality of the evidence shape the timing and scale of commitment.
Conviction remains useful when it is conditional. It can be strong without being absolute, and persistent without becoming insulated from reality. Research earns its place in a decision by improving judgment, including the judgment to revise a view.
Source notes
Closed to new investors.