One group, different situations

Six groups compress many ordered pairs

Fix a 27-position mansion ordering and preserve the distinction between the reference person and the target person. There are 27 × 27 = 729 ordered position pairs. In the reciprocal grouping derived from the three-nine sequence, each reference has one Ming target, two Ye–Tai targets and six targets in each of the other four groups.

That yields 27 Ming pairs, 54 Ye–Tai pairs and 162 pairs each for Rong–Qin, Shuai–You, An–Huai and Wei–Cheng. The totals add to 729. These are cells in a structural model, not a survey of 729 relationships. No assumption has been made that birthdays, mansion assignments or encounters are uniformly distributed.

The sequence is grounded in the three-nine ordering of the Sukuyō text in the Taishō canon (T1299); the 27-name condition used here is consistent with the calendar account of the National Astronomical Observatory of Japan (NAOJ). The counting is an explicit derivation under those assumptions. It cannot establish the probability that a reader will meet someone in a particular group, and it says nothing about a group's success rate.

This compression has a practical consequence: many different inputs receive the same group name. A category cannot retain all the information about the people, their tasks and their circumstances. More detailed language about the category does not recreate information that was never entered.

Hold the label constant and change the problem

Consider two fictional pairs preparing a neighbourhood newsletter. Pair A has trouble maintaining a weekly schedule because the participants are available in different weeks. Pair B can meet regularly but has not agreed how to acknowledge individual writing and editing contributions. Suppose both are classified An–Huai. The shared label does not make their immediate problems the same.

Pair A may need a different publication frequency or a smaller issue. Pair B needs an explicit discussion of credit and responsibilities. Those possibilities arise from the supplied conditions, not from interpreting An or Huai as a fixed personality.

Open the full-size diagram

Original controlled comparison using invented labels and situations, not observed predictive results.

The two rows of the figure preserve the same category while showing different evidence fields. They do not present two inevitable outcomes. They show why an explanation that ends with identical advice about attraction, tension and boundaries may fail to address either task.

Even matching directional offsets do not recover the missing context

For a controlled position exercise, let the first pair be Mao (昴) and Shen (参), and the second Bi (畢) and Jing (井). In the adopted 27-position sequence, which omits the Ox mansion Niu (牛), both forward offsets are three and both reverse offsets are twenty-four. Both therefore produce the same An–Huai pairing in the same directional arrangement. These are assumed mansion labels, not calculated birth records.

The matching offsets do not tell us which weeks the newsletter volunteers are free or what acknowledgement they want. Adding a near/middle/far field would still not supply those facts. Distance categories explain another positional convention, not an alternative way to obtain the participants' schedules or consent.

The same limit appears over time. One pair may work comfortably on a short project and struggle with a recurring commitment. If the birth labels remain unchanged, a report should still recognise the changed workload. Reusing the earlier advice simply because the category is stable ignores the new evidence.

Make the shared classification testable with a small transformation. Number Mao zero, then Bi one, Zui (觜) two, Shen three and Jing four in the adopted order. Mao–Shen runs from zero to three; Bi–Jing runs from one to four. Advancing both endpoints by one preserves their difference of three. The mansion names have changed, but the directed offset has not.

Now change only one input: replace Mao with Bi while keeping Shen as the target. Zero-to-three becomes one-to-three, giving forward offset two and reverse twenty-five, Shuai–You. This counterexample identifies what preserved the earlier classification and what changes it. Neither calculation has introduced availability or acknowledgement data, so a new group cannot itself resolve either newsletter problem.

Test a report by hiding its category name

Put two readings side by side and temporarily hide the names and labels. Can you identify which paragraph concerns irregular availability and which concerns attribution? If both begin with strong attraction, move to inevitable friction and conclude with generic communication advice, the text may be a category description presented as a personalised analysis.

A statement specific to Pair A should point to its schedule constraint. A statement specific to Pair B should identify the unresolved credit agreement. Neither needs to invent a hidden motive. If the available information does not establish a cause, preserve that uncertainty rather than filling the paragraph with a plausible story.

This test also catches superficial variation. Changing a few adjectives or reversing the order of a generic advice list does not add evidence. The question is whether the recommendation follows from a fact that distinguishes this situation from the other one.

Leave unknown facts unknown

If all that is supplied is a pair of mansion names, it is reasonable to explain how the classification was obtained and describe the scope of the tradition. It is not reasonable to fabricate an everyday conflict so that the reading appears deep. Ask what practical question is being considered before assigning a personalised conclusion to it.

Once concrete facts are available, they need not all be translated back into symbolism. A weekly schedule can be unworkable because of availability. An attribution dispute can remain an unresolved agreement. Both can be discussed accurately without declaring one person destined to be controlling or inattentive.

The 729-cell calculation is a structural exercise, and the newsletter cases are original comparisons. Neither is an empirical test of Sukuyō. A useful ongoing record keeps the inputs, current conditions, initial advice and later observations separate. Recording cases explains how to preserve those differences rather than collecting interchangeable stories under one label.