How YiInsight Creates a Reading: Cast, Sources, and AI Boundaries

Overview

YiInsight separates the facts used to form a reading from the later interpretive layer. In the reviewed local implementation, the cast is calculated before any AI interpretation begins; the cast snapshot and source pack are then persisted before an interpretation attempt is created. The result is a clear boundary: AI does not choose the cast, and it is not meant to turn a reflective interpretation into a guaranteed account of future events.

This guide also distinguishes source labels from source verification. Classical source, YiInsight translation, traditional commentary, and modern interpretation are separate layers, but reviewed classical quotations and a YiInsight translation are currently unavailable in the reviewed local material.

What happens first: calculate, save, then interpret

The implemented sequence places casting ahead of interpretation. Creation first calculates the cast. Within a transaction, it persists the immutable cast and its source pack, then creates an interpretation attempt. The interpretation module subsequently loads the persisted cast snapshot and source pack.

That order matters when evaluating what an interpretation can and cannot establish. The cast facts are already saved when interpretation begins, so interpretation is not the mechanism that selects them. This is a description of reviewed local implementation facts only; it does not confirm production behavior, deployed availability, or an external audit.

How the three-coin calculation fixes each line

The local three-coin calculation maps each coin as follows:

• Tails map to bit 0 and value 2.
• Heads map to bit 1 and value 3.
• A total of 6 is a changing yin line that results in yang.
• A total of 7 is a stable yang line.
• A total of 8 is a stable yin line.
• A total of 9 is a changing yang line that results in yin.

Six throws of three coins form the hexagram. The first throw sets the bottom line, and the sixth throw sets the top line. When there are no changing lines, the relating hexagram ID is null. These are local implementation mappings, not evidence of traditional-source authority for the procedure.

Why retries reuse the original cast instead of recalculating it

After creation, the local implementation saves an immutable cast and source pack for that reading. A retry creates or reuses an interpretation attempt on the same reading; it never recalculates the cast.

The expected consequence is consistency in the cast facts across retries. If an interpretation attempt is retried, that does not mean AI has selected a different cast or changed the original line structure. The safe way to assess a retry is to distinguish the fixed saved cast from the separate interpretation attempt that uses it.

Which source layers are distinct, and which are currently unavailable

YiInsight distinguishes four named layers:

• Classical source is separate from commentary and interpretation.
• YiInsight translation is a distinct layer, not a substitute label for imported text.
• Traditional commentary is separately labeled from classical entries.
• Modern interpretation is the reflective layer, distinct from source records and cast facts.

The current imported Chinese transcription is marked imported_unreviewed. It lacks edition and page information and has no independent editorial verification. YiInsight translation is null, so it is unavailable. No reviewed classical quotation and no YiInsight translation may be supplied from the reviewed local material. A missing review record, edition detail, page reference, quotation, or translation should remain visibly unavailable rather than being filled with invented content.

How AI interpretation is constrained by the saved evidence

The interpretation module loads the persisted cast snapshot and source pack. It has no casting imports, which separates interpretation from cast calculation in the reviewed local implementation.

Its output evidence IDs must exist in the source pack. Invalid evidence IDs fail rather than being accepted. The implementation also checks safety before generation and checks unsafe output before saving. These are meaningful boundaries on the described process, but they are not a claim of an audited live guarantee. They also do not convert unreviewed imported material into verified classical evidence or make an interpretation factually certain.

Use modern reflection as interpretation, not a prediction

Modern interpretation is a reflective layer, not the calculated cast itself and not a verified classical quotation. It should not be treated as a guarantee of future outcomes or as factual certainty about what will happen.

A useful reading practice is to record the fixed inputs separately: the line results, their bottom-to-top order, whether lines are changing, and the stated source status. Then keep the observation separate from the interpretation assigned to it. If a reflection cannot be connected to a valid evidence ID in the saved source pack, the described process does not support presenting it as sourced evidence.

What this method description does and does not establish

This account is limited to reviewed local implementation facts. It does not establish deployed availability, live production behavior, an external audit, or classical authority for the coin procedure or imported transcription.

It also does not make unavailable material available. The reviewed local material does not provide a reviewed classical quotation, edition details, page references, independent editorial verification of the imported Chinese transcription, or a YiInsight translation. Where those materials are absent, the appropriate conclusion is that the limitation remains open. Where an evidence ID is invalid or missing from the source pack, the interpretation process rejects it rather than treating it as supported.

Key takeaways

• The cast is calculated before AI interpretation begins.
• Six three-coin throws are recorded from the bottom line to the top line.
• The immutable saved cast and source pack are reused on retry rather than recalculated.
• Classical source, YiInsight translation, traditional commentary, and modern interpretation are distinct layers.
• The imported Chinese transcription is unreviewed, and YiInsight translation is currently unavailable.
• Modern reflection is interpretive and non-predictive, not a guarantee of future outcomes.

Conclusion

In the reviewed local implementation, YiInsight fixes the three-coin cast before interpretation, saves the cast and source pack, and reuses that cast on retry. AI is limited to interpreting persisted material with valid evidence IDs. Before treating a reflection as source-supported, check the source-layer label and note that reviewed classical quotations and a YiInsight translation are unavailable.

Does AI choose or recalculate the YiInsight cast?

No. The reviewed local implementation calculates the cast before interpretation, and the interpretation module loads the persisted cast snapshot rather than casting.

What do totals 6, 7, 8, and 9 mean in the local three-coin implementation?

Six is changing yin to yang, seven is stable yang, eight is stable yin, and nine is changing yang to yin.

Why does a retry use the same cast?

A retry creates or reuses an interpretation attempt on the same reading and never recalculates the saved immutable cast.

Which YiInsight source layers are unavailable?

YiInsight translation is null and unavailable. No reviewed classical quotation or YiInsight translation may be supplied from the reviewed local material.

Can modern interpretation predict future outcomes?

No. Modern interpretation is a reflective layer and should not be treated as a guarantee or factual certainty about future events.

Does this guide verify live production behavior or classical authority?

No. The supplied evidence describes reviewed local implementation facts only, not production deployment, an external audit, or classical authority.