Auditing Exam-Luck Claims

From the angle of corpus size, the public engine does not export an exam-result predictor. It can output four pillars, strength, pattern, shensha, and luck-cycle fields. Any education claim would require a defined outcome, a dated prediction made before the result, comparison cases, and an error count. The engine cannot claim the celebrity corpus proves educational outcomes or invent a sample statistic. These are directly checkable facts from the packet. The relevant page provides the chart output fields, but no predictor of exam luck. This limitation is a verified negative capability boundary.

Publish-Gated Sample: Field-by-Field Verification

A publish-gated sample must define each field’s evidentiary role before output. The engine’s four pillars, strength, pattern, shensha, and luck-cycle fields are descriptive, not predictive. Verification requires a dated prediction naming the exam outcome, a comparison cohort, and an error count. Any sample that omits these elements fails as an education analysis. For instance, claiming a pattern indicates exam luck without a pre-registered outcome and control cases is unsupported. The celebrity corpus cannot demonstrate educational outcomes, and no sample statistic may be invented. A compliant audit would show how each field was matched to an actual result and where the engine’s boundary lies. For foundational context on how these fields are defined, see the relevant page.

Birth-hour coverage is the first audit point for any education claim. Without a reliable hour pillar, the analysis cannot verify the hour stem-branch pair, the daymaster's strength against the hour element, or the shensha fields that some schools link to study temperament. The public engine can output four pillars, strength, pattern, shensha, and luck-cycle fields, but it does not export an exam-result predictor. Therefore, a valid education claim must define the outcome (e.g., a specific exam score or admission), record a dated prediction made before the result, list comparison cases with identical birth-hour coverage, and count prediction errors. A missing or estimated hour breaks this procedure: the hour pillar is a required input for any boundary case. Without it, no education inference can be falsified. See relevant page for hour-pillar input rules.

To audit source-field coverage, begin by checking whether the engine returns the four pillars, strength, pattern, shensha, and luck-cycle fields. Each field must be traceable to a defined educational outcome, a dated prediction made before the result, comparison cases, and an error count. If any field lacks one of these anchors, the claim is not verifiable. For example, a strength field alone cannot show exam luck unless a dated prediction and error count accompany it. See relevant page for how strength is framed. The public engine does not export an exam-result predictor, so any education analysis must stop at field coverage and cannot assert outcomes or infer from the celebrity corpus.

To audit any education claim, distinguish a sample from a population. The packet’s allowed outputs—four pillars, strength, pattern, shensha, luck-cycle fields—are per-person data points, not group statistics. A verification procedure must state: (1) the defined educational outcome (e.g., passing a named exam); (2) a dated prediction made before the result; (3) a comparison sample of cases with the same BaZi features; (4) an error count against the outcome. Without these, a single chart is anecdote, not evidence. The celebrity corpus cannot be treated as a population for educational outcomes, because no result-predictor is exported. For field-level checks, see relevant page.

Auditing Exam-Luck Claims: Questions the Census Can Answer

Any education claim from the public engine must pass a field-by-field audit. First, define the outcome: a dated exam result, not a trait like “studious.” Second, locate the prediction: it must be issued before the result, using only four pillars, strength, pattern, shensha, and luck-cycle fields. Third, list comparison cases: if the claim cites a celebrity corpus, show how each case maps to a recorded education outcome—without that, the corpus proves nothing. Fourth, count errors: every mismatch between predicted and actual results is an error; every vague or post-hoc statement is an error. A claim that cannot produce these four items is outside the engine’s negative capability. For structural context on the fields involved, see the relevant page.

Questions It Cannot Answer

A rigorous education audit must first define what the engine actually outputs: four pillars, strength, pattern, shensha, and luck-cycle fields. None of these are exam results. To verify any education claim, one would need a defined outcome (e.g., a specific test score or admission), a dated prediction made before the result, comparison cases, and an error count. The public engine provides no exam-result predictor, so no field-by-field procedure can link a pillar combination to a passing grade. Even if a celebrity chart appears to show academic success, that corpus cannot prove educational outcomes because it is not a controlled sample. Each field—year, month, day, hour—maps to a si zhu, but none carries a falsifiable study-direction claim. Without a pre-registered metric, the question “will this student excel?” remains outside the engine’s boundary.

Method replication demands a field-by-field audit trail. First, isolate the stated outcome: a named exam, a dated score, and a pass/fail boundary. Second, locate the prediction timestamp—published before results, not retrofitted. Third, extract the engine's outputs: four pillars, daymaster strength, pattern, shensha, and luck cycles. Fourth, compare each output against the claimed prediction; any mismatch voids the case. Fifth, count false positives and negatives across all comparison cases; without an error count, no replication is possible. Boundary case: if a field is missing or unverifiable, the education claim is unfalsifiable. The procedure cannot infer educational outcomes from the celebrity corpus; that would be a category error. No sample statistic may be invented. Only the listed fields are admissible evidence.

Counting Definition: A Verification Procedure

To audit any education claim from the public engine, first define the outcome in countable terms: pass/fail, rank percentile, or admission status. Then require a dated prediction made before the result, not after. For each field—four pillars, strength, pattern, shensha, luck-cycle—ask: does it state a testable educational event, and does it specify a time window? Count comparison cases: how many charts predicted success and failed? How many predicted failure and succeeded? Without such tallies, no exam luck can be verified. The celebrity corpus contains no educational outcomes; it cannot be used as evidence. Thus, the engine’s output must be treated as a boundary case: fields exist, but no exam-result predictor is available. Only an error count makes the claim auditable.

Auditing Exam-Luck Claims: Data-Quality Limits

To verify any education claim, each field must be independently audited. The four pillars, strength, pattern, shensha, and luck-cycle fields are exported, but none provide an exam-result predictor. A defined outcome must be stated before analysis, not inferred from the chart. A dated prediction made before the result is mandatory; after-the-fact matches are invalid. Comparison cases require identical outcome definitions and a pre-registered error count. Without these, the engine's outputs cannot substantiate educational performance. The celebrity corpus is not a benchmark for educational outcomes. Any sample statistic is outside the packet and must be rejected. Verification is thus a boundary case: only procedural compliance—not chart content—can be assessed.

How to Fact-Check Exam-Luck Claims

Any claim that BaZi predicts exam luck must undergo a reader-facing fact-check. First, define the outcome: a specific exam, date, and pass/fail threshold. Second, verify a dated prediction made before the result—not a retrofitted narrative. Third, compare multiple cases with the same four pillars, strength, pattern, shensha, and luck-cycle fields; record every mismatch. Fourth, count errors: a single failed prediction invalidates a universal rule. The public engine exports only those fields; it does not output an exam-result predictor. Therefore, no sample statistic may be invented, and the celebrity corpus cannot prove educational outcomes. If any step fails, the prediction remains unverified.## Auditing Exam-Luck Claims: Conclusion Limits

To audit an education claim, proceed field by field. First, a defined outcome is required: what exactly is predicted—admission, grade, ranking, or completion? Without this, no verification is possible. Second, the prediction must be dated before the result is known; a post-hoc reading of a chart cannot be counted. Third, comparison cases are needed: charts with similar pillars, strength, pattern, shensha, or luck-cycle fields should be grouped to see if outcomes differ. Fourth, an error count must be tallied: every missed or vague statement is a failure. The engine provides only four pillars, strength, pattern, shensha, and luck-cycle fields; it does not export an exam-result predictor. Therefore, the section cannot claim the celebrity corpus proves educational outcomes or invent a sample statistic. Any claim exceeding these fields is unsupported.