# T3 — TRADE STUDY 3 RESULTS

OWNER: `02_Work/scratch/T3_trade_study_results.md`  
REVIEW: regenerate from `01_System/trade_study.py` whenever a score or a weight changes.  
Split out of `T3_trade_study_weights.md` on 2026-09-07 after independent verification V6 showed that keeping the weights and the results in one file destroyed the ordering evidence the weights file existed to carry (DEF-027). The weights file is now immutable; this file is the one that moves.

---

# RESULTS — SCORED AFTER THE WEIGHTS ABOVE WERE FIXED

Scored 2026-09-07. Scores are in `01_System/trade_study.py`, each with the figure it rests on.

| Rank | Option | Weighted score |
|---|---|---|
| 1 | **A** — Aged care entity first; NDIS entity registers in parallel and takes clients second | **3.70** |
| 2 | **C** — Support coordination only; no support workforce, no aged care | **3.55** |
| 3 | **D** — Do not start; remain employed | **3.20** |
| 4 | **B** — NDIS entity first (core supports); aged care second | **2.45** |

**Margin between first and second: 0.15.** That is close, and GOV-B6.4 says a close result is escalated, not decided.

## Sensitivity — where the answer changes hands

Each row moves one criterion's weight and takes the difference from the criterion that most opposes it.

| Criterion | Weight now | Winner changes at | Hands it to |
|---|---|---|---|
| Capital at risk before first revenue | 0.25 | 0.31 | C (3.61) |
| Contribution per client per month | 0.25 | no weight up to 0.60 changes it | — |
| Time to first revenue | 0.20 | 0.28 | C (3.63) |
| Regulatory stability over three years | 0.15 | 0.23 | D (3.52) |
| Independence from the owner's own hours | 0.15 | 0.26 | D (3.64) |

## What this means

Four of the five criteria change the winner within a weight movement Zaid could reasonably make.
The study's recommendation of option A is therefore **conditional on the weights**, and the
weights are a statement of what Zaid values, not a finding about the world.

The specific fork: **raise capital preservation from 0.25 to 0.31 and support coordination wins.**
If the answer to "what is the maximum you are willing to lose entirely" is small, option C is the
correct answer and option A is not. That is ACT-003, still unanswered.

Under **GOV-B6.4 this decision is escalated to Zaid** and recorded as DEC-007, status *Parked
with Zaid*. The project recommends A; it does not decide it.