F-001 — The instrument, published.

How to read this page

Every claim in this finding carries an evidence label. Peer-reviewed means a specific published study, cited in full. Lab hypothesis means the lab's own proposed framework — stated openly, awaiting the Index's volume. Facts about the instrument's construction are stated as what they are: a design record, not evidence. Almost everything this lab believes is, today, a lab hypothesis. This page is where that stops being a weakness and becomes the method.

1 · The question

Can the person someone repeatedly becomes under pressure be read from their answers about pressure — scored on a single engine, comparably across the rooms of a life — well enough to name the pattern, measure its depth, and eventually watch both move?

2 · Why the question matters

Because the distance between intention and action is one of the best-documented disappointments in behavioural science. People sincerely intend, and then do something else — and the something else is not random. Lab hypothesis The lab's claim is that the something else is a default: a practised pattern that runs when deliberation goes offline. If that is true, asking people about their self-image measures the wrong person. You have to ask about pressure.

3 · What we actually know

Peer-reviewed Intentions predict behaviour imperfectly. Sheeran's review of the intention–behaviour literature found the relationship real but leaky — a substantial share of people who intend to act do not, and the gap is systematic enough to study (Sheeran, 2002).

Peer-reviewed Practised behaviour becomes context-cued. Wood and Neal's account of habit describes responses learned from repeated performance in stable contexts, afterwards triggered by those contexts directly, without a mediating goal — and slow to update when goals change (Wood & Neal, 2007).

Lab hypothesis What the lab knows from its own instrument, today, is only that the instrument runs and files: people answer, the engine scores, files open, and the Index counts. At publication, the Pattern Index holds 3 closed conversations and the TBVY ledger holds 1 best file (16 October 2026). Those numbers are the entire empirical content of this finding. Everything else here is design, stated so it can be tested.

4 · The evidence: the instrument as built

The instrument is eleven adaptive conversations — ten in the worst-version house (TWVY), one in the best-version house (TBVY) — running on shared mechanics: answers given as choices or free text, each answer reflected back before the next question, free text scored by keyword map, every path scored on the same engine.

Worst Mode scores eight patterns — the Builder, the Overthinker, the Avoider, the Performer, the Comfort Seeker, the Drifter, the Self-Doubter, the Seeker. The file names the winning pattern and a depth score: the winner's score as a percentage of the maximum that pattern could have scored in that conversation, reported on a clamped 5–99 scale. (Earlier site copy described the scale as 0–100; the code has always clamped to 5–99, and the copy is corrected with this finding. The floor exists because a depth of 0 would claim absence when the answers only show quiet; the ceiling because no person is entirely one pattern.)

Best Mode — The Peak, model Topaz 1.0 — scores six peak drivers (body state, deep work, people, place, challenge, rest) and files a different document: a Peak Profile built from the taker's own strongest answers, and The Prescription, three to five practices assembled from the top drivers' rule sets. It exists because the lab's first external reader — a mentor of the founder — took the instrument and asked the obvious question: it told me what runs me at my worst; what do I do to be my best? Diagnosis had a door; direction did not. Now it does.

Registers. A conversation can be taken in three registers of question language — Straight, Standard, Deep — same nodes, same scoring, register stamped on the file and the Index row. The Mirror carries all three; The Peak carries all three; the remaining nine conversations currently run Standard while their registers are calibrated. The lab publishes this incompleteness because a stamped register that was never spoken would be a false record.

Recording. Each closed conversation files one anonymous result: pattern or top driver, depth or driver scores, which conversation, the register, the model version, and a coarse country — country only, resolved at filing, never an address, never a name. The model register is public (see Methodology), and models never change silently: a result stamped 1.0 must mean the same thing in ten years.

5 · Competing explanations

Lab hypothesis The eight patterns may be the wrong joints. They are an editorial taxonomy — carved by the lab from its books and field observation — not kinds discovered in data. The Index may eventually argue for seven, or eleven; the taxonomy is versioned so it can lose that argument in public.

Lab hypothesis The answers may be performance. The instrument asks about pressure rather than self-image precisely to narrow the gap between the presented self and the acting one — but a determined presenter can perform in any register. Nothing in a single sitting rules this out.

Lab hypothesis Depth may measure extremity of answering rather than spread of pattern — a person who picks the strongest option everywhere scores deep everywhere. The normalisation (share of the pattern's own maximum) corrects for conversation length, not for answer style.

6 · What remains uncertain

Validity against established instruments: untested. Test–retest stability: unmeasured. Whether depth predicts anything about a life over time: unknown. Whether the taxonomy travels across cultures: the Index is designed to find out, and has not yet. The lab lists these on the agenda rather than in the small print, because they are the work.

7 · What a person can test

The Fieldwork path: take the Mirror, read the file, take the Peak, run the Prescription for two weeks, and re-take the Mirror in thirty days. The protocol is deliberately small — one pattern, one prescription, one re-test — because a test a person will not actually run produces no evidence at all.

8 · How we would know whether it worked

At the person's scale: depth movement between the two files, in the direction the Prescription aimed at. At the lab's scale: re-test rows accumulating in the Index until depth-before and depth-after can be compared honestly, and — the validation studies on the agenda — the engine's reads checked against instruments the field already trusts. If depth does not move when the conditions change, the thesis is wrong in a way the lab will publish.

9 · The new questions

Does a Work default predict a Love default (F-002)? Do dominant defaults differ across borders, and along what lines (F-003)? Does a peak profile — the conditions of a person's best days — predict which prescription they will still be running in a month? The instrument was built to make these askable. This paper is the lab keeping its first promise: a lab that asks for trust publishes its instrument first.

References

Sheeran, P. (2002). Intention–behavior relations: A conceptual and empirical review. European Review of Social Psychology, 12(1), 1–36.

Wood, W., & Neal, D. T. (2007). A new look at habits and the habit–goal interface. Psychological Review, 114(4), 843–863.

Finding F-001 · Institute for Default Research · Published 16 October 2026. Counts stated as of publication; the live ledgers are on the Pattern Index. Corrections to this finding will be appended, never silently edited.