R0021/2026-03-25/Q003 — ACH Matrix¶
Matrix¶
| H1: Predominantly quantifiable | H2: Predominantly subjective | H3: Mixed | |
|---|---|---|---|
| SRC01-E01: OpenAI 6 strategies, 0 quantifiable | -- | ++ | - |
| SRC02-E01: Anthropic 7 recommendations, 1 semi-quantifiable | -- | ++ | + |
| SRC03-E01: Google 7 recommendations, 3 quantifiable | - | + | ++ |
| SRC04-E01: Microsoft 5 recommendations, 0 quantifiable, "art not science" | -- | ++ | - |
Legend:
- ++ Strongly supports
- + Supports
- -- Strongly contradicts
- - Contradicts
- N/A Not applicable to this hypothesis
Diagnosticity Analysis¶
Most Diagnostic Evidence¶
| Evidence ID | Why Diagnostic |
|---|---|
| SRC04-E01 | Microsoft's explicit statement that prompting is "more of an art than a science" is the most diagnostic evidence because it is a vendor's own characterization of the practice they sell. It strongly discriminates between H1 (quantifiable) and H2 (subjective). |
Least Diagnostic Evidence¶
| Evidence ID | Why Non-Diagnostic |
|---|---|
| SRC03-E01 | Google's inclusion of some quantifiable data points (temperature, accuracy claims) makes it the least diagnostic — it supports both H2 (majority qualitative) and H3 (some quantifiable). |
Outcome¶
Hypothesis supported: H2 — Vendor guidance is predominantly subjective. Approximately 85-90% of recommendations across all four vendors use qualitative language without measurable criteria.
Hypotheses eliminated: H1 — No vendor provides predominantly quantifiable guidance.
Hypotheses inconclusive: H3 — Partially supported as a refinement of H2. A small minority of recommendations (primarily from Google) are quantifiable.