# T1 — TRADE STUDY 1: WHEN TO REGISTER

OWNER: `02_Work/scratch/T1_trade_study_weights.md`
REVIEW: superseded, never edited, if a criterion or a weight changes.

## WHY THIS FILE EXISTS AT ALL

`01_System/trade_study.py` and DEC-004 both cited a weights record at this path. Independent
verification pass V9 found the path did not resolve — the ordering evidence for trade study 1
lived only inside the decision register's own free text, which is the party that made the
decision asserting its own procedure. That is DEF-047.

This file is written after the fact and **says so**. It is not backdated and it does not pretend
to be contemporaneous evidence.

## THE WEIGHTS, AND WHEN THEY WERE SET

The weights below were recorded in DEC-004 on **2026-08-20T09:40 AEST**, and the scores at
**11:05 AEST** the same day. The one hour twenty-five minute gap between the two register fields
is the only ordering evidence trade study 1 has, and it is a claim by the builder, not a proof.
GOV-C3.2 gives a builder's statement about its own work zero weight, so what follows is stated as
what it is.

| # | Criterion | Weight | Why it carries this weight |
|---|---|---|---|
| 1 | Capital at risk before first revenue | 0.30 | Zaid is the sole capital provider and the registration window is four to six months of fixed cost with no revenue. It is the criterion where being wrong is unrecoverable. |
| 2 | Addressable market during the registration window | 0.25 | Agency-managed participants are unreachable without a certificate. How much of the market that removes decides whether trading early is worth anything. |
| 3 | Regulatory durability past July 2027 | 0.25 | Mandatory registration reaches personal care and daily living supports in July 2027 (SRC-015). An option that does not clear that date is not an option. |
| 4 | Operational burden while the application runs | 0.20 | A one-person business preparing an audit and serving clients at the same time is the practical constraint, not the theoretical one. |

Total: 1.00

## WHAT V8 AND V9 FOUND, AND WHAT CHANGED

**V8 (2026-09-07).** The published totals — 2.85, 4.05 and 2.30 — could not be produced from the
published matrix. The correct totals under these weights are **3.50, 4.10 and 3.25**; one was out
by 1.20 on a five-point scale. The published sensitivity ("A2 leads by 1.20; flips above about 55
per cent on addressable market") was wrong on both figures: the margin is **0.60** and the flip is
at **37 per cent**, taken from capital at risk. The scores were hard-coded strings in the study
builder while trade studies 2 and 3 were computed.

**V9 (2026-09-07).** The arithmetic was fixed but the study's own list of alternatives still
described the option that had been replaced, and DEC-004 still published the wrong sensitivity in
three delivered artefacts. Both are now corrected.

**The third alternative changed.** It was "remain unregistered indefinitely", which the study's own
evidence shows is unlawful for a core-supports business from July 2027 — so the three-alternative
bar GOV-B6.1 sets was being met by padding it with an option that could not be adopted. It is now
"do not trade core supports before the certificate; use the window to run support coordination and
build referrals", which is lawful, credible, and which the model can actually cost. The eliminated
option is still shown, as eliminated, with its reason.

## THE HONEST LIMIT OF THIS RECORD

The ordering for trade study 1 is attested, not proven, exactly as it is for trade study 3
(DEF-027). The protection that does not require trusting the author is the same one: the weights
are live in the workbook's `TradeStudy3` sheet for trade study 3, and for all three studies the
matrix, the weights and the figure behind every score are published in the delivered study, so a
reader can redo the arithmetic in about two minutes. Checker C43 redoes it on every build — which
is what nobody did for three verification passes.