The Ten Gods in Bazi: What 1,285 Real Charts Actually Show
Two charts, one god, two lives
Here are two charts from the same teaching corpus, and they share one feature: 伤官 (Hurting Officer) sits in plain view in both.
The first is a woman born with the pillars 壬寅 庚戌 己丑 庚午. Her day master is 己 earth. The month stem and the hour stem are both 庚 metal — and metal is what earth produces, so both are 伤官. Two of them, exposed. The note beside this chart records her occupation in four characters: a vice-president at a listed company. And here is the telling part — the note's own explanation for how she reached that floor never once points at 伤官. It points at 寅, the officer on the guest side, feeding her own 午 (her 禄, the seat of authority): "the officer on the outside grows into the power on her side, so she holds office." It even explains why she is the vice and not the chair — a clash between her day branch 丑 and her hour branch 午 (丑午之穿).
The second is a man born 壬子 戊申 辛未 丙申. His day master is 辛 metal, and water is what metal produces, so the year stem 壬 is 伤官. The hour stem is 丙 fire, and fire presses metal — that is 正官 (Direct Officer). Textbook readers look at this and see 伤官见官, "Hurting Officer meets Officer," one of the most-quoted warnings in the entire system. The note on this chart records one phrase: 光棍汉 — a man who never married. But again, the note's own reasoning for that outcome never mentions the officer. It reads his spouse palace instead: the day branch 未 is a wealth vault, but the vault is sealed; with no wealth star in the chart, the spouse is represented by 伤官 and 食神, yet the 食神 子 (his year branch) pierces that palace (子水穿妻宫). The marriage is blocked at the level of the palace, not the ten gods.
Same god — 伤官, sitting in plain view in both. One reached the executive floor of a listed company; the other never married. And in neither case did the teacher explain the outcome by naming the god. Both readings happen at the level of palace, position, and clash — exactly the level a presence test throws away. So the honest question is not "what does 伤官 mean." It is whether the label carried any of the explanation at all.
Ten gods are verbs, not verdicts
The ten gods are not ten personality types, and they are not ten fortunes. They are ten relationships between one character — your day master — and every other character in the chart. Each relationship is a verb.
Five verbs, to be exact: something feeds me (生我), I feed something (我生), something presses me (克我), I press something (我克), or something stands where I stand (同我). Split each by yin and yang and you have ten.
This is why "good god / bad god" is the wrong axis. There is no good verb. Pressing is excellent if you are shaping steel and terrible if someone is pressing you. Output is wonderful if the world is buying and ruinous if the world is watching. The verb does not carry the verdict. The sentence does — and the sentence is the whole chart.
The ten, and what each one is doing
同我 — 比肩 and 劫财. The people standing where you stand. Sharing, competing, splitting the bill, and also the only ones who will show up when the room clears.
我生 — 食神 and 伤官. What you put out. Skill, output, speech, appetite, the thing you make that did not exist before. 食神 makes steadily; 伤官 makes sharply and does not check who is in the room.
我克 — 正财 and 偏财. What you can get your hands on and convert. Money, yes, but also anything you manage, run, or turn into something else. 正财 is the salary; 偏财 is the deal.
克我 — 正官 and 七杀. What presses on you. Structure, bosses, law, reputation, the cost of being seen. 正官 is a rule you can live inside. 七杀 is pressure that does not negotiate.
生我 — 正印 and 偏印. What feeds you. Schooling, backing, shelter, the ability to absorb a hit and keep going. 正印 is the scholarship; 偏印 is the thing you taught yourself at 2am.
One caution on the older translations: you will find charts that call 正官 a woman's husband and 正财 a man's wife. Those are conventions from a specific time and place, not mechanics. The mechanics are the verb.
Why the slogans sound so convincing
伤官见官, 为祸百端 — "Hurting Officer meets Officer, and trouble comes from every side." It is a beautiful line. It scans. It is memorable enough to survive being wrong.
And it is right often enough to survive on anecdote alone. Which is exactly the problem, because a practitioner's case book is made of anecdotes. Every chart in it was chosen because it was worth talking about. A book of cases is not a sample of lives; it is a sample of stories worth telling.
So we stopped quoting the slogan and tested it.
What we did: 1,285 charts, 64 tests
We keep a regression corpus for this kind of question: 1,285 charts extracted from 97 files of 段建业's published teaching notes (the blind-school tradition, 盲派), deduplicated from 4,557 raw extractions.
Every chart had to pass four checks before it was admitted: the characters are real sexagenary pairs, the yin-yang polarity is internally consistent, the month pillar is consistent with 五虎遁, and the hour pillar is consistent with 五鼠遁. That last pair is the load-bearing one — a random or OCR-mangled four-character string almost never satisfies both 遁 rules at once, so a chart that passes is very likely parsed correctly.
Then we took eight of the most-quoted 十神 pairings — 伤官见官, 食神制杀, 伤官生财, 杀印相生, 伤官配印, 官印相生, 财官相生, 比劫夺财 — and counted a pairing as present only when both gods appear among the four visible heavenly stems. That 透干 rule matters: a blind-school reader reads what is exposed, not what is buried. Each pairing was tested against eight outcome tags. Eight pairings times eight tags is 64 tests.
The outcome tags are extracted from the notes' own factual sentences — the ones opening with "实际 / 果然 / 结果 / 后来" — so they mean "what the notes say happened," not an independently audited biography. That is a real caveat and we are keeping it in the open, because a practitioner writes up the cases that came out interestingly.
What came back: nothing survived
The strongest result in all 64 tests was 伤官生财 — output feeding wealth — against marriage breakdown: 22.3% (31 of 139 charts) versus a 13.1% baseline. Lift 1.70, z = 2.95.
Second was 伤官配印 against an official career: 21.5% versus 14.2%, z = 2.25.
With 64 tests run at once, the bar for "probably not noise" sits around |z| ≥ 3.4. Nothing cleared it. Not one of the eight pairings.
And the headline slogan itself? 伤官见官, n = 78 charts, against legal trouble: 8% versus 10% baseline. Against marriage breakdown: 19% versus 14%. Against an official career: 18% versus 15%. The most feared combination in the system is, in this corpus, unremarkable.
We also re-ran everything on the 225 charts whose labels come only from strictly factual sentences rather than the practitioner's framing. There, nearly every cell drops below 25 charts and there is nothing to say at all. Absence of evidence is not evidence of absence — but it does mean nobody should be quoting these slogans with confidence.
The one finding that survived — and how fast it broke
A wider scan of the same corpus, run over 97 structural facts against the outcome tags, produced 1,533 tests and left exactly one survivor above the |z| ≥ 3.5 bar: 财星入墓 — the wealth star sitting inside one of the four tomb branches 辰戌丑未 — against 发财, acquiring wealth. 22% versus 13%, lift 1.73, z = 3.80.
It is an appealing result because it fits the blind-school idea of a 库 as a vault: the money is stored, sealed, and it takes luck-pillar pressure to crack it open. 段建业 returns to 墓库做功 repeatedly, and this looked like the data agreeing with him.
Then we redefined 财库 the way a reader might from a short description — any tomb branch whose hidden stems include a wealth star — and ran the same test again. The result collapsed to z = 0.90. Nothing.
One definition clears a multiple-comparison correction. A second, equally reasonable definition of the same idea finds no signal at all. That gap is the actual finding, and it is worth more than the positive result was.
How the ten gods are actually read
段建业's own answer to all of this is already on the record, and it is the answer the data points at: 做功 is not "this god is present." It is who is doing work on whom, from where, and with what strength behind them.
A presence test throws away three things, and those three things are most of the reading.
Position — 位置. The same god in the month pillar and in the hour pillar are not the same event. The month is the house you grew up in and the current of your working life; the hour is what you are like when nobody is watching.
Host and guest — 宾主. Day and hour are your side. Year and month are the world's side. A god standing on your side and the identical god standing out there describe completely different situations.
Strength — 旺衰. A 正官 with roots in its branch and a 正官 floating alone with nothing under it are not the same pressure. One is a job with a contract; the other is a title with no budget.
This is also where our two charts at the top diverge — and where the slogans fall apart. Both have 伤官 prominent, but the teacher never uses it to explain either life. Hers is read through 寅 feeding her 午 禄 and a 丑午 clash that makes her a deputy. His is read through a sealed wealth vault in the spouse palace and a piercing 食神 that blocks the marriage. Same visible god. Completely different sentences — written in palace and clash, not in the ten-god label.
Check it on your own chart
One: get your four pillars. If your birth time is rounded or uncertain, fix the hour pillar first — a quarter of the chart hangs on it.
Two: find your day master. It is the heavenly stem of the day pillar, and it is the "I" that every other character is measured against.
Three: for each of the other three stems, ask one question only — does it feed me, do I feed it, does it press me, do I press it, or is it the same as me? Answer that before you reach for any name.
Four: split by yin and yang to get the god's name. Same element, opposite polarity, different god.
Five: now do the part that actually carries the reading. Note where each god sits, and whether it has roots — a matching hidden stem in its branch. Write the verb next to the pillar, not the verdict. "Output, doubled, exposed, officer buried" is a sentence you can test. "Bad god" is not.
Then pick whichever god appears twice in your chart and check it against something you have already lived through. Backwards first. That is the only order in which you can tell whether any of this is working.
Frequently asked questions
Are the ten gods the same as planets in Western astrology?
No. Planets are objects with positions; the ten gods are relations computed from your day master, and there is no fixed object behind them. The same star is your 正官 in one chart and your 七杀 in another, because the relation changed, not the star.
Is 伤官 always bad?
No. In 段建业's own corpus the strongest statistical signal involving 伤官 pointed towards marriage breakdown, and it still failed the correction bar. 伤官 is output. Whether output costs you depends entirely on what is receiving it and where it sits.
Why do different sites give me different ten gods?
The day master is rarely the disagreement. What differs is whether a site counts hidden stems at all, how it weights them, and whether it labels 偏印 as 枭神. Differences almost always come from hidden-stem weighting, not from your chart.
Does this mean bazi does not work?
No. It means the slogan version does not. We tested the crudest possible reading — "god X is present, therefore outcome Y" — which is not how any serious reader reads a chart. Failing that test is what the tradition itself would predict, and 段建业 says so explicitly.
Where do the 1,285 charts come from?
From 段建业's published teaching notes in the blind-school (盲派) tradition — 97 source files, deduplicated from 4,557 raw extractions. We store only the four pillars, gender, derived labels and the source filename. We do not store the original commentary.