Zi Wei Year Pillar: Lunar New Year or 立春? (The 13-Day Window)
zi wei dou shu and bazi split their year boundaries differently — and within one app the setting alone can redraw the whole chart. This page pins the 2026 window to the date and shows the same birth charted under both conventions.
Direct answer
In Zi Wei Dou Shu the birth-year stem-branch is taken by the lunar new year in mainstream practice (this site's setting, and the default in most practitioner software), while iztro's own default is 立春. For 2026 the two boundaries are 17 February (lunar new year) and 4 February 04:02:08 (立春): a birth on 10 February therefore charts as 乙巳 under the lunar-new-year convention and 丙午 under the 立春 convention — and because the year stem drives the birth-year four transformations, the entire 12-palace chart differs, not just one pillar.
Key terms in this page
- 年界 (year boundary)(nián jiè)
- — which calendar instant turns the birth-year stem-branch; this decides the year's four transformations.
- 正月初一 (lunar new year)(zhēng yuè chū yī)
- — first day of the first lunar month; the mainstream Zi Wei year boundary.
- 立春 (start of spring)(lì chūn)
- — the first solar term; the BaZi year boundary and iztro's default for Zi Wei.
- 生年四化 (birth-year transformations)(shēng nián sì huà)
- — four star transformations derived from the year stem; a different year stem changes which stars transform.
01What the year boundary actually controls
In Zi Wei the year stem is not just a label: it selects the birth-year four transformations (禄/权/科/忌) and, together with the lunar month and hour, fixes the location of the life palace and the body palace. Change the year stem and those derivations move — which is why a boundary difference in this system is larger than in BaZi, where a year change usually moves one pillar.
The two conventions exist for historical reasons. Classical Zi Wei texts count the year from the lunar new year; 立春 belongs to the solar-term reckoning that BaZi uses. Software has inherited both, and which one a product uses is often invisible unless you compare two apps on a birth inside the window.
02Worked example: 10 February 2026, both conventions
With iztro 2.5.3 at 06:00 on 10 February 2026 (Asia/Shanghai), the lunar-new-year convention returns four pillars 乙巳 庚寅 乙卯 and a chart built on year stem 乙; the 立春 convention returns 丙午 庚寅 乙卯 and a chart built on year stem 丙. The day and month pillars are identical; the hour branch is identical. Everything that depends on the year stem — the four transformations above all — changes.
The near misses pin the boundary: 3 February 2026 returns 乙巳 under both conventions (neither boundary has passed), and 18 February returns 丙午 under both (both have passed). The disagreement is exactly the thirteen days between 4 February 04:02 and 17 February — the same window that splits BaZi charts, but with a larger downstream effect because the year stem feeds the transformation table.
A second near miss is the 立春 instant itself: a birth at 04:01 on 4 February is still in the previous year under the 立春 convention; one at 04:03 is already in the new one. Software that rounds the term to the day will get one of those two wrong.
03What this site does, and how to check another app
This site sets the year boundary to lunar new year for both the natal chart and the horoscope layers (yearDivide and horoscopeDivide), matching mainstream practitioner practice and the convention our cross-validation against the Renhuai dataset confirmed. The chart's convention panel states the setting, and charts born inside a dispute window carry an extra banner naming the other convention's year stem-branch.
To check another app: open a birth inside the window (for 2026, any date from 4 to 16 February), then compare the year stem-branch printed on the chart. If one app says 乙巳 and the other 丙午, the difference is this setting, not a calculation error.
04Limits
This page does not claim that one boundary is correct for Zi Wei. Both are in documented use; the mainstream convention is lunar new year, and that is what this site states rather than hides.
It also does not resolve the interaction between the two boundary systems: a product that uses 立春 for Zi Wei is internally consistent, but a reading that mixes a BaZi-style year with a Zi Wei-style month is not a convention — it is a bug.
05Reader checklist
- The birth date was checked against the year's lunar new year and 立春 instants
- The year stem-branch on the chart was compared between the two conventions
- If inside the window, the four transformations were compared, not just the year label
- The app's documented year-boundary setting was checked
FAQ
Which boundary is mainstream for Zi Wei?
Lunar new year (正月初一) is the mainstream practitioner convention and this site's setting. iztro's own default is 立春, which is why some software differs on window births.
Why does the whole chart change instead of one pillar?
The year stem selects the birth-year four transformations and participates in locating the life and body palaces, so changing it moves many derived placements at once.
My app and this site disagree on a February birth — which is right?
Both follow their configured convention. Check the date against the 13-day window; if it falls inside, the difference is the boundary setting.
Does the horoscope (流年) layer use the same boundary?
It should. This site sets both natal and horoscope boundaries to the same convention; products that set only one produce charts whose current-year layer drifts from the natal chart.
What if I was born exactly at 立春?
Then the instant matters: before 04:02:08 on 4 February 2026 the 立春 convention still reads the previous year. A birth record accurate only to the hour cannot resolve this minute.
Sources reviewed on 2026-10-31
- iztro 2.5.3 (yearDivide normal vs exact, 2026-02-03/02-10/02-18) — computed 2026-10-31; snapshot archived with the page
- lunar-javascript 1.7.7 (立春 2026-02-04 04:02:08; lunar new year 2026-02-17) — engine output
- Our Zi Wei engine (divination-core + iztro, yearDivide=normal) — cross-validated against a 78,864-chart dataset at 100% under the same convention
Computed values on this page were produced by running the engine versions listed above at the reviewed date; school disagreements are reported as disagreements, not resolved to a single answer.
Related pages
- The Four Transformations Disagree on Three Stems: 戊, 庚 and 壬The ten-stem four-transformation table looks fixed until you compare editions: three stems differ between the standard table and the Zhongzhou (Wang Tingzhi) table. This page shows the exact rows that differ and what changes on a chart.
2026-10-31