Keep the original Sukuyō condition

Decide whether an entry can be judged before counting it

A personal reading log may contain a precise expectation, a broad impression, an unfinished event and a sentence written after the outcome. Those entries do not become comparable merely because each has a mansion label beside it. Before counting apparent matches, determine what each entry actually claimed and whether there was an opportunity to observe it.

Keep not-yet-due entries distinct from missed conditions, and distinguish ambiguous wording from a clearly observed result. A log can improve the honesty of recollection without becoming a study of predictive validity. This article offers a recording practice, not a validated scoring system for Sukuyō.

Preserve the input and the original sentence

Store the method used to obtain the mansion labels and identify the comparison: two people's birth references, or one birth reference against a selected day's mansion. T1299's three-nine discussions distinguish these subjects. That textual distinction helps prevent a later review from quietly replacing one comparison with another.

Preserve the sentence written at the time, its date, scope and intended deadline. Put later explanations in new entries rather than editing them into the original. If the first note simply said that cooperation would improve, a later detailed account of a specific event should not be presented as though the detail was predicted in advance.

The US National Archives' document-analysis method asks readers to consider creator, audience, time, purpose and supporting material. We adapt those general questions to a personal record. The institution has not supplied a Sukuyō experiment here, and its worksheet does not certify the conclusions of a diary.

Consider Bi as the reference. In the cited three-nine order, Zui and Shen intervene before Jing, making Jing inclusive position four, An. If Jing is another person’s birth mansion, the target is that person. If Jing is the mansion assigned to a selected day, the target is that date. The label An matches, but the target field does not. One observation of a conversation cannot silently be used to assess both comparisons.

Follow one invitation through a deadline

Imagine two people planning a neighborhood walk. Before sending an invitation, the recorder writes: “By Wednesday evening, both participants will have confirmed the meeting place and starting time.” This is an invented, observable expectation, not a forecast calculated from actual birth data.

The recorder then sends a proposal. One participant confirms both details. The other agrees to the location but says the starting time is still uncertain. At the Wednesday deadline, the original condition has not been met. A reply is not the same as confirmation of both specified details.

Open the full-size diagram

Original planning-log example, not a measured prediction or a real participant record.

If the second participant confirms on Thursday, add that event with its date. It may be a successful resolution of the plan, but it does not retroactively meet the Wednesday condition. The distinction is not punitive. It protects the meaning of the original sentence so that it can be reviewed fairly.

Now change only the criterion. Criterion A is “both people replied”; criterion B is “both confirmed the place and time.” One complete confirmation and one reply with an uncertain time meet A but not B. The record must use the criterion written before the invitation, rather than selecting the easier sentence after the replies arrive.

Your actions belong in the outcome record too

Sending a clear proposal, offering an alternative time or following up may make agreement more likely. If those actions occurred, they are part of the process that produced the result. A later agreement cannot simply be credited entirely to the reading while the practical work disappears from the log.

The opposite outcome also needs context. Perhaps the plan was canceled for a reason outside the recorded question, or the participants never had a chance to respond. That may affect whether the original claim can be evaluated, but the explanation should be documented rather than selected after the fact to preserve a preferred conclusion.

You can record that a reading prompted useful reflection or helped you organize a question. That is a claim about its role in your activity. It is different from showing that it accurately predicted an external event. Keep those possible benefits and claims in separate fields instead of forcing them into one accuracy score.

Keep the full sequence available for later comparison

Do not retain only the most striking apparent matches. Keep unresolved entries and observations that did not fit, with enough context to understand their status. Nor should every unobservable case be marked wrong simply to make a spreadsheet complete. A meaningful review starts by preserving the limits of each record.

This exercise specifies no sampling frame, control group or statistical test, so it reports no percentage of effectiveness. Its narrower purpose is to expose changed wording, missing observations and selective recollection. For comparisons across different practical situations, see same group, different dynamics.

When discussing real cases with others, separate the shareable method notes from unnecessary identifying details. Use participant codes where possible, keep sensitive birth inputs separately if they are needed for recalculation, and do not publish private messages merely to make the case vivid. If details are omitted, label the omission rather than replacing them with invented facts.

A final status can remain open: “Wednesday condition not met; one time confirmation missing; revised arrangement awaiting reply.” That is a usable record, not an unfinished piece of storytelling that needs a reassuring ending. For documenting the source and calculation version behind a log, see checking Sukuyō sources.