Investment decisions are made with incomplete information, and no method removes that. The choice is between deciding as though the future is knowable and deciding in a way that accounts for it not being knowable.
Probabilistic thinking is the second option.
A forecast is not a plan
A prediction states what will happen. A plan describes what to do across a range of outcomes, and how much to commit to each.
The difference becomes clear under a single test: what would have to be true for this decision to be wrong, and what would that cost? A decision that cannot answer that question is not a decision — it is a conviction.
Thinking in ranges changes behaviour in three concrete ways:
- It makes the cost of being wrong part of the decision, not an afterthought.
- It makes position size a deliberate choice rather than a consequence of enthusiasm.
- It makes reassessment a scheduled activity rather than a crisis response.
Base rates before narratives
When assessing a new situation, the most useful first step is usually to ask what typically happens to things of this kind, before considering what makes this one special.
The reason is that a vivid, specific story — about a company, a technology, or a market — feels more informative than a general statistic, even when it is less reliable. Starting from the base rate and then adjusting for what is genuinely different is more disciplined than starting from the story and adjusting for the base rate.
Scenarios and what they are for
A scenario is a coherent description of a possible future and its consequences for the portfolio. Scenarios are not predictions; their value comes from being considered together.
A useful set has three properties:
- The scenarios are genuinely different in their implications, not variations on one outcome. If every scenario leads to the same action, the exercise has not tested anything.
- They are specific about transmission. Not "recession" but how a recession reaches this portfolio: which revenues fall, which spreads widen, which positions become difficult to sell.
- They include the uncomfortable one. The scenario that is least likely and most damaging is exactly the one that is easiest to leave out.
Eonvera's Scenario & Causal Intelligence module exists for this purpose: evaluating portfolio outcomes and capital allocation choices under alternative market paths, through probabilistic research, stress testing, and multi-scenario analysis — rather than resting on a single forecast.
Sizing: the decision that carries the risk
Two positions can be equally well researched and carry very different consequences. What differs is size.
The questions that should set size:
- How much does this cost if the thesis fails entirely? Not "if it falls 20%" — if it goes to zero, or if the issuer defaults.
- How correlated is it with what is already held? A small position in something that fails alongside everything else is not small.
- How liquid is it if the portfolio needs cash? Size should respect what could actually be sold in the conditions when selling would be necessary.
- What is the exit condition? Decided in advance, in writing, while the position is still comfortable.
The discipline of changing your mind
The hardest part is not forming a probabilistic view. It is revising it when the evidence changes, without revising it merely because the position is losing money.
The two are distinguishable only if the conditions were written down beforehand. "I would reduce this if credit spreads widened beyond a certain level" is a rule. "I feel worse about this than I did" is not.
This is the principle behind Eonvera's stated investment philosophy: adaptation matters more than prediction. The aim is not to be right about the future, but to make decisions that remain defensible across a range of futures — and to recognise quickly when the evidence has moved.