Verification data
What each claim is, how far it matches, what it proves and what it cannot prove — all stated. We dare show this table to experts.
Benchmark 1 · 560/560 full-field comparison
Against a 518,000-chart sample library, compared field by field across 8 dimensions — 560 items, zero mismatches:
- · 14 major stars × 12 palaces
- · 12 minor stars + Lucun & Tianma
- · Ming/Body palace branches
- · Five Element Bureau
- · Zi Wei star's palace
- · Si Hua (Lu, Quan, Ke, Ji)
- · Decade start/end
- · Annual GanZhi / monthly flow
Honest note: the sample library is itself generated by the same engine, so this benchmark is “internal consistency self-evidence” — it proves our engine configuration (leap months, hour boundaries, gender handling) is error-free, and is not by itself an external correctness argument.
Benchmark 2 · 32 public famous cases re-run under two schools
We took the 32 public famous cases with complete birth hours, charted each under both the standard and Zhongzhou algorithms, and cross-compared: Ming Palace majors, minors, Si Hua and all 12 palaces' majors agree 100%. Differences appear only in four categories of auxiliary stars — Ming Zhu/Shen Zhu, Jiekong, Tianshi/Tianshang, Dahao/Suipo — exactly where the two schools genuinely diverge, matching theoretical expectation perfectly.
Honest note: the birth hours of public historical figures are mostly later estimates, so this benchmark likewise validates internal consistency rather than serving as a strict external argument.
Benchmark 3 · Disputed stems against both classical tables (strongest evidence)
The disagreement between the two mainstream classics (Zi Wei Dou Shu Quanji / Quanshu) concentrates on two Heavenly Stems: Geng and Ren. We verified our default table (Quanji edition = mainstream convention) against the Quanshu edition, item by item, using the two-book comparison of the classical “Formula for Placing Lu-Quan-Ke-Ji” cited on Wikipedia — every difference is correctly implemented and reproducibly switchable:
| Stem | Quanji edition (default) | Quanshu edition |
|---|---|---|
| 庚 | 太阳禄/武曲权/太阴科/天同忌 | 太阳禄/武曲权/天同科/太阴忌 |
| 壬 | 天梁禄/紫微权/左辅科/武曲忌 | 天梁禄/紫微权/天府科/武曲忌 |
| 戊 | 贪狼禄/太阴权/右弼科/天机忌 | 贪狼禄/太阴权/右弼科/天机忌(both books agree) |
The other 8 stems agree in both books (for Wu, a few versions give Taiyang Ke; both mainstream books give Youbi Ke). See Si Hua schools.
Strength of evidence at a glance
| Claim | Strength |
|---|---|
| Engine implementation includes all known edge-case fixes | Strong (fact) |
| 3 disputed stems (Geng/Ren/Wu) checked against an independent source | Strong (independent check) |
| 560/560 full-field comparison | Medium (internal consistency self-evidence) |
| 32 famous cases re-run under two schools, 100% agreement | Medium (internal consistency) |
| Star-placement rules implemented against mainstream classics, independently checkable | Medium (checkable rules) |
Our boundaries (stated as well)
- · We do not support the special placement methods of certain more niche schools (e.g. the “Lianzhen-Pojun in Mao” convention).
- · Charting is deterministic computation — reproducible and checkable; interpretation belongs to cultural research, and a correct chart does not make predictions certain.