Does Qimen Work? Evidence, Reproducibility, and Limits
Qimen calculations can be made reproducible when the time, offset, boundary rules, method, and engine version are fixed.
Overview
Qimen calculations can be made reproducible when the time, offset, boundary rules, method, and engine version are fixed. That reproducibility proves consistency of the declared calculation; it does not by itself prove that symbolic interpretations predict external outcomes. Evaluate three layers separately: mechanism validation checks whether the board follows the formula; source validation checks whether a stated rule has a locatable authority; outcome validation would require preregistered predictions, comparison baselines, and measured results.
At a glance
- Page role
- Foundation concept
- Terms covered
- Qimen accuracy · Qimen evidence
- Verification focus
- Evaluate three layers separately: mechanism validation checks whether the board follows the formula; source validation checks whether a stated rule has a locatable authority; outcome validation would require preregistered predictions, comparison baselines, and measured results.
- Stop rule
- Do not report confidence as a probability unless it was statistically calibrated. Avoid guaranteed dates, diagnoses, legal conclusions, financial returns, mind-reading, and claims that lack a chart-specific evidence ID.
Direct answer and scope
Qimen calculations can be made reproducible when the time, offset, boundary rules, method, and engine version are fixed. That reproducibility proves consistency of the declared calculation; it does not by itself prove that symbolic interpretations predict external outcomes.
- Names and related terms: Qimen accuracy
- Names and related terms: Qimen evidence
How the rule is formed or checked
Evaluate three layers separately: mechanism validation checks whether the board follows the formula; source validation checks whether a stated rule has a locatable authority; outcome validation would require preregistered predictions, comparison baselines, and measured results.
Keep the cast moment, UTC offset, method version, palace, and interacting symbols visible before adopting an interpretation.
How to use this in a reading
Use a reading as a structured hypothesis generator: name the chart fact, propose more than one explanation, identify a disconfirming observation, and choose a low-cost next check. This makes the advice falsifiable without claiming certainty.
- Name the visible chart fact before assigning symbolic meaning.
- Compare the asker and matter anchors with every layer in their palaces.
- Record at least one supporting condition and one counter-signal.
- End with a reversible action and an observable review point.
Worked contrast
A test suite that reproduces 2,196 charts across a year is strong evidence that one engine implements its stated rules consistently. It is not a clinical trial, investment backtest, or proof that readings improve decisions.
Common mistake and limit
Do not report confidence as a probability unless it was statistically calibrated. Avoid guaranteed dates, diagnoses, legal conclusions, financial returns, mind-reading, and claims that lack a chart-specific evidence ID.
Sources and editorial basis
- 猪野毅「奇門遁甲の基礎的研究」Abstract and pp. 161–185: history and vocabulary; nine palaces; Three Wonders and Six Instruments; Eight Gates, Nine Stars, and deities; 排宮 and 飛宮 method distinction.A 2010 Hokkaido University research article used for historical vocabulary, the basic board structure, and the documented distinction between rotating-palace and flying-palace approaches. It does not test predictive accuracy.
- OpenFate Qimen hour-chart methodOpenFate docs/QIMEN_SYSTEM.md, “Professional calculation contract,” “Interpretation and evidence contract,” and “Rule-source map,” source-reviewed 2026-09-12.Product-specific reference for OpenFate’s fixed-offset civil-time, hour-Qimen, rotating-plate, Chaibu implementation and its deterministic chart fields. A source review on 2026-09-12 does not prove live deployment, other schools, or predictive validity.