How investing actually works
Why invest by rules rather than predictions
A prediction says what will happen. A decision method says what you do, under stated conditions, whatever happens. The difference shows up at the only moment that matters — when you have to act and the outcome is still unknown.
7 min de lectura · Actualizado el 2026-09-07
El memo
A strategy is not a forecast about the market. It is a set of decisions made in advance, stating what you do under which conditions — so that the decision does not have to be invented in the moment.
Para recordar
- A prediction says what will happen. A method says what you do.
- The rule is written before the moment, which is the only time it can be written calmly.
- Judge a decision by its method, not by how the one instance turned out.
La distinción
Being right and having a method are different things, and neither implies the other. A guess can be right; a method can be wrong on a given occasion. What separates them is that only one of the two can be examined before you find out.
El error frecuente
Treating a rule as a claim about the future. A rule does not assert that the condition it watches will occur, nor that acting on it will pay. It asserts only what you do if it occurs — which is why it can be written down without knowing anything about what comes next.
En resumen — You cannot decide in advance what the market will do. You can decide in advance what you will do — and that is the only one of the two that is actually available to you.
Markets have fallen fifteen per cent over six weeks. Commentary is available in unlimited quantity, and it disagrees with itself. What do you do now?
That question has a property worth noticing: no forecast answers it. Even a forecast that turns out to be correct does not tell you what to do, because it arrives without a size, without a date, and without any statement of what happens if it is wrong. The gap between knowing what might happen and knowing what to do is where most investment decisions are actually made — and it is the gap a strategy is supposed to close.
Two different objects
A prediction is a claim about the world: this will rise, that is overvalued, the fall has further to run. It can be true or false, and you find out afterwards.
A decision method is a claim about you: under these conditions, this is what I do. It is not true or false — it is followed or not followed, and you can check which at any moment.
The distinction sounds like a technicality, and everything else here follows from it. A prediction is answerable to how things turn out — a point forecast to the extent it proves accurate, a probabilistic one to the extent it proves well calibrated over many occasions. That second form is worth pausing on, because it is the more defensible kind: a forecast that gave a fall thirty per cent odds is not refuted by the fall arriving, and it can be genuinely useful without ever being right about a single occasion. What it cannot escape is that its usefulness is settled somewhere in the future. A method's is not: it does not assert anything about the future at all, so it assigns a response to a condition rather than a likelihood to an outcome. That is why a method can be written down in full while knowing nothing about what comes next, and a prediction cannot.
It also changes what can be examined. A prediction is judged once, after the fact, by an outcome that mixes the reasoning with luck in unknown proportions. A method can be read before the fact by someone who did not write it — including by you, later, when the outcome is known and the story has become tempting to retell.
Why this is a choice, and not a finding
It is often said that investors fail for lack of method rather than lack of intelligence. Stated as a fact about people, that is more than the evidence supports: skill, information and judgement plainly differ between investors, and sometimes decisively.
It is better understood as a principle for deciding — and the argument for it does not depend on any claim about who fails or why.
Consider what each approach asks of you. Relying on judgement requires you to be right, repeatedly, in conditions designed to make that difficult: incomplete information, time pressure, and an outcome that will not tell you afterwards whether your reasoning was sound or you were fortunate. Relying on a method requires you to specify the response in advance, once, calmly — and then to keep applying it.
The second may still do worse. It is simply a demand you can meet deliberately, whereas being reliably right is not something anyone can undertake to do. Choosing a method is a decision about which kind of demand you are willing to take on, not a claim that ability is irrelevant.
What a rule costs
A method that is only ever described in its favour is being sold rather than explained, so it is worth being plain about the price.
A rule will sometimes act when judgement would have been better, and sit still when judgement would have moved. It cannot recognise a situation nobody anticipated, because it only knows the conditions it was given. And it will occasionally look foolish for months at a time, which is not a malfunction — it is the same property that lets it act without your agreement in the other direction.
There is a subtler cost. A rule you will not actually follow is worse than no rule at all, because it supplies the feeling of having a method while the decisions are still being made in the moment. The usefulness of a rule is not a property of the rule alone; it depends on whether the person holding it can live with what it does.
What makes a rule a rule
Three properties, and the third is the one usually missing.
It states a condition precisely enough that two people reading it would agree on whether it has been met. "If markets look weak" is not a condition; it is a mood with a threshold implied and unstated.
It states the response, including how much moves. An approach that says what to watch but not how far to go has left the difficult half undecided.
And it exists in writing, before the situation arises. This is the part that quietly disappears. A rule recalled from memory during a fall is not a rule; it is a preference, and it will be recalled in whichever version is easiest to act on that day. The writing is not bureaucracy. It is what makes the rule the same object on a calm day and a bad one.
The mistake this is meant to prevent
The habitual error is to judge a decision by how it turned out.
It is an understandable one, because the outcome is the only part that is easy to see. But a single outcome is one draw from a distribution: a decision with no reasoning behind it can be rewarded, and a sound decision can be punished, and neither result says much about what to do next time. Judging by outcomes reliably teaches the wrong lesson — most sharply when the result was good, because then there is nothing prompting a second look.
Judging by method asks a different question: was the decision specified in advance, was it applied as written, and would it be applied the same way again? Those can be answered whatever the market did. They are also, unlike the outcome, entirely within reach at the moment you have to act.
Compruébalo
Two people sell everything on the same day of a market fall. One had written, months earlier, that they would reduce exposure if a stated condition was met; it was met that week. The other sold because the news had become unbearable. A year later both are equally better off. Did they do the same thing?
Ver la respuesta
No, and the identical outcome is what makes the case worth thinking about. One decision can be described, repeated, and examined by someone else: the condition existed before the day did, so it is possible to ask whether it was applied correctly and what it would do next time. The other cannot be repeated, because nothing about it was specified — the same person facing the same ambiguity next year has nothing to apply. The shared outcome tells you about that year, not about either decision. Judging by the result would rate the two equally, which is precisely the reasoning this distinction exists to interrupt: an outcome is one draw, and one draw does not separate method from luck.
Preguntas frecuentes
- Isn't a rule just a prediction in disguise?
- It looks that way, and the difference is real. A rule of the form "if this measurement crosses that level, reduce this exposure" makes no claim that the crossing will occur, or that reducing will turn out well. It fixes a response to a condition. A prediction fixes the condition itself. The test is simple: a rule can be written down completely without knowing anything about the future, and a prediction cannot.
- If rules are better, why does anyone use judgement?
- Because judgement can do things a rule cannot — recognise a situation nobody anticipated, weigh circumstances no measurement captures. The argument here is not that judgement is inferior; it is that judgement leaves nothing behind. It cannot be handed to someone else, checked before the fact, or applied identically on a bad day. Many serious investors use both, and the useful question is which parts of a decision are which.
- What actually makes something a rule rather than an intention?
- Three things, and the last one is the one people skip. It states a condition specific enough that two people would agree on whether it has been met. It states the response, including how much moves. And it exists in writing before the situation arises — because a rule recalled from memory during a fall is indistinguishable from a preference, and it will be remembered in whichever form is most comfortable at the time.
- Doesn't following a rule mean I stop thinking?
- The opposite, if anything: the thinking moves earlier, to a moment when it can be done calmly and revised without cost. What the rule removes is the need to invent a response under pressure, with incomplete information and an unpleasant screen in front of you. Deciding to keep following a rule through a stretch where it looks wrong is itself a decision, taken repeatedly — and it is where most systematic approaches are actually abandoned.