These eight cover where online bazi content goes wrong most often: the year switch line, the midnight rule, solar time, daylight saving, match percentages, strength presuppositions, hidden stems, and the zodiac standing in for the self. Each entry carries an engine-computed counterexample — not opinion against opinion, but a fact you can verify by hand. Convention numbers (R numbers) point to the matching cards on the methodology page.
Many sites replace the whole chart with the zodiac animal of the birth year, then hand out personality and fortune by animal. The core of bazi is the day master — the stem of the day pillar; the year branch is only one of four pillars.
| Where it breaks (computed counterexample) | Reading the animal by the calendar year gets it wrong: born 1990-01-15 12:00, the year pillar is 己巳 Ji Si (Snake), not Horse — lichun 1990 (1990-02-04 10:14) had not arrived. The day master comes from the day pillar: that day is 庚辰 Geng Chen, a Geng Metal day master. |
|---|---|
| What the engine does | The engine switches the year pillar and zodiac at the lichun instant (R30) and takes the day master from the day stem (R25). |
| Year pillar at 1990-01-15 12:00 | 己巳 Ji Si (Snake, before lichun) |
| Day pillar at the same instant | 庚辰 Geng Chen (day master = Geng Metal) |
The other common error: switching the animal and year pillar at the Spring Festival or on January 1. The switch line is lichun, an astronomical instant, and it differs from both — lichun can fall on February 3, earlier than the February 4 many sites assume.
| Where it breaks (computed counterexample) | Lichun 2025 was February 3, 22:10, while the Spring Festival (January 29) was long past. Born 2025-02-03 21:00, the year pillar is still 甲辰 Jia Chen (Dragon); only at 2025-02-04 01:00 does it become 乙巳 Yi Si (Snake). Cutting at the festival marks both as Snake; cutting at February 4 marks the late evening of the 3rd as Snake too. Both are wrong. |
|---|---|
| What the engine does | The engine uses an offline-generated lichun table for 1900-2100 (R30); atypical years like 2025-02-03 22:10 are directly lookable. |
| 2025-02-03 21:00 (one hour before lichun) | Year pillar 甲辰 Jia Chen (Dragon) |
| 2025-02-04 01:00 (after lichun) | Year pillar 乙巳 Yi Si (Snake) |
For births in 23:00-24:00 (late Zi) two schools disagree on the day pillar: the mainstream Zi Ping school rolls it into the next day, another school changes the day at midnight. Many calculators silently pick one, and users never learn which rule produced their chart.
| Where it breaks (computed counterexample) | Born 1991-03-23 23:30: the mainstream late-Zi school gives day pillar 癸巳 Gui Si, the midnight school 壬辰 Ren Chen — a whole pillar apart. The month pillar (Xin Mao) is identical. The engine defaults to late-Zi-rolls-forward (R3) but computes both and says so (R10). |
|---|---|
| What the engine does | sect=1 (default, late Zi rolls into the next day) and sect=2 (midnight change) are both published (R3/R10). |
| 1991-03-23 23:30 · late-Zi school (default) | Day pillar 癸巳 Gui Si |
| 1991-03-23 23:30 · midnight school | Day pillar 壬辰 Ren Chen |
Hour boundaries follow local solar time; the same clock reading corresponds to different sun positions at different longitudes. A calculator with no birth-place field gives Urumqi and Shanghai the same hour pillar at 12:55.
| Where it breaks (computed counterexample) | 1995-06-01 12:55: at longitude 119°E the hour pillar is 戊午 Wu Wu (Wu hour); at 80°E the local sun runs about two hours behind and the pillar flips to 丁巳 Ding Si (Si hour). The closer a birth sits to an hour boundary, the more this correction rewrites the whole pillar. The engine applies the true-solar correction from the entered longitude (R4) and documents how the central meridian is picked. |
|---|---|
| What the engine does | True solar time v1 = longitude term of ±4 minutes per degree (R4); the ±16-minute equation-of-time limit is published on the methodology page. |
| 1995-06-01 12:55 · 119°E | Hour pillar 戊午 Wu Wu (Wu) |
| 1995-06-01 12:55 · 80°E | Hour pillar 丁巳 Ding Si (Si) |
China ran daylight saving time from 1986 to 1991; during those six years the clock ran one hour ahead of standard time. Most calculators ignore this, which shifts every hour pillar in the window by a whole hour.
| Where it breaks (computed counterexample) | Born 1988-07-01 11:30: the engine subtracts the DST hour and computes 10:30, giving 乙巳 Yi Si (Si hour); a DST-blind calculator reads 11:30 straight and lands on the Wu hour — the no-DST control, 1993-07-01 11:30, is indeed 戊午 Wu Wu. Same clock reading, a whole branch apart (R5). |
|---|---|
| What the engine does | The CN_DST 1986-1991 rough boundary is declared in the engine as an approximation (R5), with its edge-date limits published. |
| 1988-07-01 11:30 (engine minus DST → 10:30) | Hour pillar 乙巳 Yi Si (Si) |
| 1993-07-01 11:30 (no-DST control) | Hour pillar 戊午 Wu Wu (Wu) |
Zodiac match and synastry sites routinely output a percentage precise to the digit. Those numbers have no checkable derivation: change the weighting and the verdict changes, and the weighting is never published.
| Where it breaks (computed counterexample) | The twelve zodiac-compatibility pages on this site refuse to output percentages: they give structural relations (six harmony, trine, clash, harm, break, punishment) with overlap declarations, each checkable against the relation tables. The compatibility calculator outputs checkable structural facts, not scores. |
|---|---|
| What the engine does | Standing declaration: no compatibility percentages (zodiac-compatibility pages and the stance section of the methodology page). |
| What the pairing pages give | Structural relations + overlap declarations |
| What they refuse to give | Percentage scores — numbers with unpublished weights cannot be checked |
Verdicts like too much wealth for a weak self presuppose a quantified strength reading first; only then does each god get a direction. Skipping strength and declaring more killings is bad is the shared pattern of every marketing page.
| Where it breaks (computed counterexample) | The engine makes strength a published tally (R20): the element counts across the five relations and the thresholds (strong at 0.55, weak at 0.45, R21) are printed on the favorable-element page, with both school readings side by side (R22). The same god acts in opposite directions under weak versus strong structures — that is a computed result on the page, not a slogan. |
|---|---|
| What the engine does | Strength = published tally (R20), declared thresholds (R21), two schools both computed (R22). |
| Strong threshold | 0.55 or above (declared, R21) |
| Weak threshold | 0.45 or below (declared, R21) |
| Schools | Both strength readings computed and shown (R22) |
Every earthly branch hides one to three stems, and the main qi is only the first of them. Reading Yin as nothing but Wood discards the other two tenants inside it — exactly where many shen sha and ten-god judgements land.
| Where it breaks (computed counterexample) | Yin hides 甲 Jia, 丙 Bing and 戊 Wu; Xu hides 戊 Wu, 辛 Xin and 丁 Ding. The branch pages and the twelve-stages pages list each branch's full hidden-stem table and ten-god attribution, computed — not one animal and one line of main qi. |
|---|---|
| What the engine does | The hidden-stem table is copied verbatim from the engine table (same practice as R24); branch pages compute each branch's tenants and their gods. |
| Hidden stems of Yin | 甲 Jia · 丙 Bing · 戊 Wu (main qi 甲 Jia) |
| Hidden stems of Xu | 戊 Wu · 辛 Xin · 丁 Ding (main qi 戊 Wu) |
Every enumerable component of ba zi has its own page here. Each page is computed by the same engine and can be checked item by item — none of it is a personality table.