Pinned before, not after

Ask anyone why they own something and you will get a good answer. Ask them a year later, after the position has worked, and you will get a different good answer — one that fits what happened.

This is not dishonesty. Memory is reconstructive. The reasons that turned out to matter are easier to recall than the ones that did not, so the remembered thesis quietly converges on the outcome. By the time you are asked, you are not retrieving the original argument. You are assembling a plausible one from what you now know.

Which means the most common way an investment thesis fails is not that it turns out wrong. It is that it stops being available to be wrong.

What the derivation date does

When Teniq derives a reason a company is worth owning, it stamps the date and the document. The reason is written down at the moment it is derived, from the filing it came from, and it is not editable afterward.

That is the whole mechanism. It is unglamorous and it is the part that makes everything downstream mean anything.

A reason with a derivation date can be tested. You can look at what was claimed in March and what the filings said in June, September and December, and the comparison is real because the March version is fixed. A reason without a date cannot be tested against anything, because it will have moved.

The difference it makes to a track record

Almost every claim of research quality in this industry is unfalsifiable by construction. The argument is described after the fact, the supporting evidence is selected after the fact, and the reader has no way to establish what was actually believed beforehand.

Pinning inverts that. The corpus accumulates a dated record of arguments that existed before their outcomes did. Some of them held. Some weakened. Some were tested by a filing within a quarter. All of it is on the record with dates attached, which means it can be counted rather than characterized.

This is also why we can publish our own numbers rather than describing our process. You cannot report how many reasons survived a year unless you know, unambiguously, what the reasons were a year ago.

What pinning is not

It is not a prediction. A pinned reason is a statement about what a filing supports, not a forecast about what will happen. A reason identifying owned distribution as the source of a cost advantage is an argument about a structure. It is not a view on the share price, and nothing in the system treats it as one.

It is also not permanent. Reasons weaken. Reasons get superseded. The pin does not preserve the reason's status, it preserves its original text and date so that its status can be tracked honestly. What holds today is a live field. What was claimed in March is history, and history does not get edited.

The uncomfortable part

Pinning means being wrong in a way that stays visible.

An argument that was reasonable on the evidence available and then fell apart does not disappear. It sits in the record with its derivation date, next to the filing that undermined it. There is no mechanism for quietly removing it and no version of the corpus where it did not happen.

That is the cost, and it is the point. A system that could edit its own history would be more flattering and would tell you nothing. The value of the record comes entirely from the fact that we cannot go back and improve it.