Building a Repeatable Process

Scope, and this article carries the tightest constraint in the pillar. It supplies no template, no checklist, no worked example of a process, and nothing that terminates in a decision or an allocation. A published process is a system, and this portal does not supply systems. It also does not claim that having a process improves outcomes — the pillar's governing finding is that awareness of an effect does little to remove it, and a process presented as the remedy would claim an efficacy the evidence does not support. What the article does is state what a process can actually do, which is narrower than usual and better supported. This is the final article of Pillar 32 and closes it.

The eight problems were the same problem

Every one of those is a failure of evidence about oneself, and specifically the absence of a contemporaneous record. In each case the reader is reasoning competently from what is available to them — and what is available is a reconstruction assembled after the outcome was known. Memory supplies the closed positions and not the open ones. A streak supplies the sequence and not the reasoning. A decline supplies the feeling and not the prior intention. A published gain supplies the result and not the entry. Nobody in any of those situations is being careless. They are all trying to learn from a record that was never made. That is the single defect this pillar has been circling, and it is the only one a process addresses.

What a process can do

A process does not improve judgement, and nothing in the evidence reported in this pillar suggests it does. What it does is make judgement inspectable.

Its function is to record what was believed, and why, before the outcome was known. Which is the one input every failure above lacked.

The distinction that makes this describable without becoming a system is between a process and a rule. A rule determines what is decided — buy at this level, hold for this long, allocate this proportion. A process governs how a decision is reached and what is preserved about it. The first is a system and this portal supplies none. The second is a description of practice, and it can be discussed without anyone being told what to conclude.

Properties, described rather than prescribed

Written before the outcome. A record made afterwards is a reconstruction, and it is subject to exactly the selection effects the rest of this pillar documents.

Reasons, not only actions. A list of transactions is a statement of what happened; it contains nothing about what was expected.

Dated. Without a date it is impossible to establish what was known at the time, which is the whole point.

Containing what would change the view. The falsifiability test from Reading Financial Media Critically applies to a reader's own reasoning as much as to anyone else's: a position held for reasons that no observation could contradict is not a position based on reasons.

What it cannot do

It cannot prevent a bias. The governing finding of this pillar applies to processes as much as to knowledge.

It cannot improve an outcome. A well-documented decision and a badly documented one have the same consequences; the documentation affects what can be learned, not what happens.

It cannot substitute for information, and it cannot resolve the questions Pillar 31 identifies as depending on facts about a person that no framework knows.

The conclusion of Pillar 32, stated plainly. Nothing in this pillar supports the claim that a reader can become unbiased. The evidence points the other way, repeatedly and across several decades of measurement. What the evidence does support is narrower and still worth something: a person who records what they believed and why, before knowing how it turned out, can later examine their own reasoning against a sample that was not assembled to flatter them. That is not a remedy for anything. It is the difference between being able to audit yourself and not being able to — and given that every failure in this pillar arises from reasoning about oneself on defective evidence, it is the one improvement that is actually available. You cannot become unbiased. You can become auditable.