How bazifor.me computes a reading

Ten claims, stated plainly

These are the ten things this site commits to about how its engine works and what it does not do; each can be checked against the rule ids printed in every reading.

  1. Every reading is deterministic: the same birth data or floor plan always produces the same result.
  2. Every line of output names the classical rule it applies, with a rule id a practitioner can check.
  3. Birth times are converted to true solar time using the birth city's time zone on that date, its longitude and the equation of time; Singapore and Malaysia clock times before 1982 use the UTC+7:30 offset of that era.
  4. Character strokes for naming follow the Kangxi Dictionary (康熙字典) count on the traditional form.
  5. The 81-number stroke verdicts use a two-state scale, auspicious or not, until a cited finer table is adopted.
  6. Sound-alike checks run at three strengths: same syllables and tones, same syllables with different tones, and merged consonants; severe words cap the grade.
  7. The engine does not predict events; it computes classical structures and states their traditional reading.
  8. No generative AI writes any reading; the text is rule output and hand-written guides.
  9. Workspaces live behind an unguessable link, are never listed, and are not used for anything else.
  10. Where schools disagree on an element or a stroke count, the engine records its choice and marks it as disputed.
The ten claims in Chinese (十条方法声明)
  1. 每一份解读都是确定性的：相同的出生资料或户型图，永远得到相同的结果。
  2. 每一条输出都注明所依据的古典规则及规则编号，行家可以逐条核对。
  3. 出生时间按经度与均时差换算为真太阳时；1982年前新加坡和马来西亚的钟表时间采用当时的UTC+7:30。
  4. 起名用字的笔画按康熙字典，以繁体字形计数。
  5. 八十一数理的吉凶只分吉与不吉两档，待采用有出处的更细分表后再细化。
  6. 谐音检查分三档：音节与声调相同、音节相同而声调不同、声母合并；严重的谐音词会限制评级上限。
  7. 引擎不预测事件，只计算古典结构并陈述其传统解读。
  8. 任何解读都不由生成式人工智能撰写；文字来自规则输出和人工撰写的指南。
  9. 工作区位于无法猜测的私人链接之后，从不公开列出，也不作其他用途。
  10. 各流派对五行或笔画有分歧时，引擎记录所作选择并标注为有争议。

For the full pipeline read how it works, for who stands behind it see the author, and for the rules explained see the guides.

How the six life-area scores are built: each chip's classical basis, and what the percentile beside a score is and is not.