Analysis Methodology
How Saju Works analyzes a chart, made transparent
Why publish this?
There are many online Saju services, but few explain which theories they apply or how. Saju Works documents the source classics, weighting, role of AI, calendar computation, and the limits of the analysis so that readers can examine results critically and use them with their own judgment. This is also our basic posture toward Meongri: it is a humanities tool, not divination.
Weighting of the three classics (6:3:1)
Every analysis on this site integrates the three foundational Meongri texts with the following weighting:
Japyeongjinjeon (子平眞詮) — 60%: The backbone of structural analysis. Centered on the month branch (wolryeong), it determines the eight standard patterns (jeonggwan, chilsal, jeongjae, pyeonjae, jeongin, pyeonin, sikshin, sanggwan), the special patterns (yangin, geonrok, wolgeop), and the variant and transformation patterns (jonggyeok, hwagyeok).
Gungtonbogam (窮通寶鑑) — 30%: Seasonal balance — cold/warm/dry/moist (johu). Reads which season the day master was born into, whether the chart contains the johu yongsin it needs, and how the balance is broken if it is.
Jeokcheonsu (滴天髓) — 10%: Distinctive insights such as cheong-tak-jin-ga (clarity, turbidity, truth, falsehood) and gangyu-sunyeok (rigid/yielding, going-with/against). These shape the integrated score and final commentary.
Why 6:3:1?
Japyeongjinjeon carries the structural backbone — gyeokguk and yongsin — so it gets the largest share. Gungtonbogam adds seasonal flesh to that skeleton; Jeokcheonsu lays the final layer of insight on top. The ratio also matches the priority Korean Meongri academia generally places on these three texts when teaching them.
Role and limits of AI
Saju Works implements core logic — calendar computation, gyeokguk and yongsin determination, johu analysis — as deterministic algorithms. The same input always yields the same structural output. The AI (a large language model) is used only to render that structured data as natural-language interpretation; it does not invent or alter the underlying chart structure.
- The AI has been trained on classical Meongri texts; it does not generate new schools of theory or unorthodox interpretations.
- Phrasing in the natural-language commentary may vary between runs for the same chart, but the underlying structural data (gyeokguk, yongsin, johu, etc.) does not.
- On rare occasions, the AI-generated phrasing may include inaccuracies. Treat all output as reference material rather than authoritative judgment.
How the calendar (manse) is computed
Saju Works derives the month pillar from the 24 solar terms (jeolgi). We follow the Meongri convention of starting the year at Ipchun, with each solar term boundary fixed to the minute by astronomical calculation. Hour pillars default to Korean or Japanese standard time, with an optional longitude correction (true solar time) for the birthplace — Seoul −32 minutes, Tokyo +19 minutes, and so on.
Citation and sourcing
All columns and learning material on the site are written from the source classics. Where possible, the original Chinese text is shown alongside Korean, English, and Japanese readings. Celebrity analyses use only publicly available birth information; when the birth time is not public, the limitations of hour-pillar reading are stated explicitly. When citing Korean or Chinese Meongri scholars, we identify which of the three classics the cited theory belongs to.
Limits of this analysis
- Saju is a system for analyzing temperament and broad life rhythms from birth information; it is not a deterministic prediction of the future.
- This analysis is not a basis for medical, legal, or financial decisions. Always consult a qualified professional in those domains.
- Two people with the same chart will live different lives depending on environment, education, effort, and relationships. The chart is a starting point, not a verdict.
- If the birth time is uncertain or unknown, the reliability of hour-pillar analysis drops sharply. We recommend using a time confirmed to the minute when possible.
Our commitment
Saju Works does not dress Meongri up as fortune-telling; it treats it as a branch of East Asian humanities. By stating our sources, weights, the role and limits of AI openly, we aim to be a tool that readers can trust. This page is updated whenever new theory is incorporated or the algorithm is improved.