Measured Against the Masters: How Our Consultations Align With the Classical Astrologers

Five questions worth asking any consultation

Before you trust a reading, from a person or an AI, it is worth asking where it comes from. Here are our answers.

Where does my reading come from?

Rules written directly from named classical books, by a practitioner of both traditions.

Will the same chart get the same reading tomorrow?

Yes. The engine decides what your chart holds, the same way every time.

Can I check how closely it follows the classics?

Yes. We publish dated figures, book by book, on our BaZi and astrology pages.

Will it hand me a verdict on my life?

It shows what your chart holds and why, and leaves the decisions with you.

Can I ask in my own language?

Yes. Ask in any of our 15 languages, and the answer comes back in yours.

Why this matters to you

An AI can write a confident, fluent reading for any chart. What you cannot see is whether that reading follows the tradition or the AI's own guesswork. Ask twice, and two readings of the same chart may not agree.

If you have tried an app or a reading and come away unsure, that is usually why. A reading is only as good as the method behind it, and that method is usually out of sight.

The classical books change this. Their authors wrote down their method, then showed it at work on real charts. That gives a fixed standard anyone can check against, and we hold our engine to it in public.

How a consultation works

Your consultation has two parts, and each does one job.

The engine reads your chart. It casts your chart from your exact birth details, then reads it by rules written directly from the classical books.

• In BaZi, it reads your Day Master (the element at the heart of your chart) and how strong it is, your chart's structure, the element your chart needs most, the combinations and clashes between elements, and your decade and annual phases.
• In Hellenistic astrology, it reads whether you were born by day or by night, the planets that rule each area of your life, and the classical timing methods that show which themes each year and each chapter of life bring forward.

What this means for you: the same chart always gets the same reading, and every reading traces back to a named book.

The AI explains what the engine found. It receives the engine's findings, in the order a consultant would tell them, and answers your question in plain words and in your language.

What this means for you: you get the depth of a classical consultation without having to learn its terms first, and you can keep asking until the picture is clear.

How we test it

The classical authors did more than state rules. They printed real charts and wrote down what they saw in each one.

Take a typical example. An author prints a birth chart and says it is a Wealth structure, that its Day Master is weak, and which element it needs most. That gives us three judgments to check. We run our engine on the same chart and see whether it reaches the same three.

We test on the worked charts in these books:

• BaZi: The True Interpretation of Ziping (子平真诠; Qing dynasty, 18th century), and Dripping Heavenly Marrow (滴天髓), whose worked charts come from its commentary, Dripping Heavenly Marrow Elucidated (滴天髓阐微; Qing dynasty, 19th century).
• Hellenistic astrology: Vettius Valens' Anthologies (2nd century) and Dorotheus of Sidon's Carmen Astrologicum (1st century).

Two more books guide the engine but are not scored yet. The Precious Mirror of Adversity and Success (穷通宝鉴; Qing dynasty, 19th century) records too few worked judgments of these kinds, and Ptolemy's Tetrabiblos (2nd century) works through no example charts.

How we count, and why we count strictly

Each figure in our tables is the share of the author's judgments where our engine reaches the same reading. We count strictly, because a number that flatters us is worth nothing to you.

• A different reading counts against us. So does a reading that matches only in part.
• Silence counts against us. If the author gives a reading and the engine gives none, it counts as a miss. Otherwise, an engine that stayed quiet whenever it was unsure would look better than it is.
• Only what can be tested is counted. Some charts are printed without a birth time or a planet's position, so no one can cast them exactly as the author did. These, and a few topics where our engine follows a different classical author, are set aside and listed in our records.
• Small samples are left out. A figure needs at least 10 of the author's judgments. Three out of three looks perfect and tells you nothing.
• Every table is dated. When the engine improves, we publish a new table with a new date. Old figures are never quietly edited.

How we keep ourselves honest

A test only means something if the engine cannot be quietly shaped to pass it.

• A sealed set. Part of the BaZi authors' example charts is locked away. No one opens them or runs the engine on them until a single scoring run at the end of our current round of work, and that result will be published too. Today's BaZi tables come from the charts we work with day to day; the sealed set is the stricter test. The astrology books hold too few worked charts to set some aside.
• Every change, on every chart. Each change to the engine is rerun on every test chart at once. A change that fixes one chart but breaks another is reviewed by hand, never hidden inside a better total.
• The original texts. We work from the Chinese texts of the BaZi classics, compare editions, and record every correction made to a damaged passage. For astrology, we work from scholarly translations and check doubtful passages against other ancient sources.
• The AI is tested too. Each change to how the AI writes is tried on the same questions many times over, and kept only if it stays inside what the engine found.

What we learned building it

Building an engine on books written centuries ago brought its own problems. Here is how we solved the main ones, and what each means for your reading.

• The books do not always agree. Each kind of judgment follows the book that treats it most fully, and every open choice is made once, with its reason recorded. For you: one consistent method, not a blend that shifts from answer to answer.
• An author sometimes seems to break his own rule. That usually hides a condition he states elsewhere, so we look for it before changing anything. For you: the engine keeps the author's nuances instead of flattening them.
• Old texts are terse. A rule can be a handful of characters. We read each rule against the author's own worked charts, which show what he meant in practice. For you: rules applied the way the author applied them.
• Timing is the hardest part. The books often judge a ten-year phase in a few words, and reading one means weighing it against the whole chart. It is where our figures are lowest and where our current work is focused. For you: we show you where we stand, in a dated table you can check.
• Fluent writing drifts. An AI writes smoothly, and smooth writing can slide into guesses. Ours is given only what the engine found and your question. For you: answers grounded in your own chart.

Why few can do this

Measuring readings against the classical authors takes two things that rarely come together: deep knowledge of the doctrine, and the patience to write it into an engine rule by rule and test it chart by chart.

Sheng Pin Kai was built by a practitioner of both Chinese and Western metaphysics who did exactly that. It is why we can show you our figures, and why we will keep publishing them as the engine improves.

Try it on your own chart

The best test is your own chart. Ask one question about your career, your relationships or the years ahead, and see where the answer comes from.

The Free plan has no time limit and needs no credit card. Start your consultation, and see the current figures on our BaZi and astrology pages.