刑 (Punishment) and 害 (Harm): How Bokki Actually Flags Chart Friction
Three small, fixed tables decide whether your chart's card mentions friction at all — and one pair of branches can trigger two different traditional labels for the exact same two characters.
Bokki Editorial
The short answer
Bokki's spouse-palace note (the card that reads inner and close-relationship tendencies) sometimes says a chart's branches carry no major knots, and sometimes names specific friction: 삼형, 자형, or a named 害 pair. None of this comes from reading a mood or a fate — it comes from checking your chart's four branches against three small, completely fixed lookup tables: three punishment (刑) groups, four self-punishing branches, and six harm (害) pairs. A separate fourth table, six clash (沖) pairs, is checked independently and reported on its own line. Running these tables over 5,000 charts, 75.1% show at least one 刑 or 害 pattern, and 7.9% trigger two different labels from the exact same two branches at once — the tables themselves overlap.
The rule — three tables, a membership count
刑 (punishment) has two forms. Group punishment checks whether at least two of a fixed three-branch set are present anywhere in your four pillars: {Tiger, Snake, Monkey} (인사신 삼형), {Ox, Dog, Goat} (축술미 삼형), or {Rat, Rabbit} (자묘형, a two-branch group). Self-punishment checks whether any of four specific branches — Dragon, Horse, Rooster, Pig — appears twice among your four pillars (its own duplicate is the punishment: 진진, 오오, 유유, 해해). 害 (harm) is simpler: six fixed pairs (Rat-Goat, Ox-Horse, Tiger-Snake, Rabbit-Dragon, Monkey-Pig, Rooster-Dog), and if both branches of any pair are present anywhere in your chart, that pair's label appears. None of these checks look at which pillar a branch sits in, its element, or your day master — they only ask which of the twelve branches showed up. And the tables are not independent of each other: Tiger and Snake are two of the three members of the 인사신 group, and they are also, on their own, the 인사 harm pair — so a chart with Tiger and Snake present but no Monkey earns both '인사신(寅巳申) 삼형' and '인사(寅巳) 해' for the identical two characters.
Worked examples — three charts Bokki has already used elsewhere
1) Born 15 February 1980, 12:00, Seoul (庚申 戊寅 戊午 戊午, already the worked example in Bokki's 역마 article): branches present are Monkey, Tiger, Horse, Horse. Tiger and Monkey are two of the three 인사신 members, so '인사신(寅巳申) 삼형' fires even without Snake. Horse appears twice (day and hour pillars), so '오오(午午) 자형' fires too. And Tiger-Monkey is also one of the six clash pairs, so the clash line fires as well — three separate friction/clash labels from just four branches. 2) Born 2 November 1988, 09:10, Busan (戊辰 壬戌 辛酉 癸巳, already the worked example in Bokki's 화개 article): branches are Dragon, Dog, Rooster, Snake. Rooster and Dog are the 유술 harm pair, so '유술(酉戌) 해' fires, and Dragon and Dog are a clash pair, so the clash line fires too — no punishment table matches this chart at all. 3) Born 6 May 2000, 13:20, Seoul — the exact chart Bokki's own source code uses to document a bug it once had: this chart has a clash (子↔午) but no 刑 or 害 match at all. Before a 2026-07-30 fix, Bokki's card would have told this person 'no conflict' — technically true only of 刑/害, but read by a person as covering 沖 too. The card is now split into a separate clash line and friction line, so a chart like this one states the clash plainly instead of denying it by omission.
What this calculation doesn't decide
These tables never say whether the friction they find is good or bad — Bokki reads a clash as information about movement, not a warning (D-007), and reads the absence of 刑/害 as steadiness rather than a reward. When a birth time is unknown, Bokki checks only the year, month and day branches it actually has and says so, rather than declaring an unmeasured hour branch free of friction; measured across 500 real users' charts against eleven possible birth hours (2026-07-29), every single one of the 259 charts that showed no friction on three known pillars turned out to gain a 刑 or 害 pattern for at least one of the eleven hours — a chart's 'no friction' verdict describes the pillars it can see, not a guarantee about the one it can't. Bokki also doesn't compute other schools' further refinements (whether a punished or harmed branch is itself supported or clashed by yet another branch, or severity weighted by day-master strength) — only the plain membership checks described above.
Closing note — three fixed tables, checked against 5,000 charts
In short: whether your chart's relationship card mentions friction comes down to three small, unchanging tables checked against the twelve branches your four pillars actually contain — no mood, no day-master weighting, just membership. Across 5,000 charts, 56.5% show a punishment pattern, 42.8% show a harm pattern, 75.1% show at least one of the two, and 7.9% show both a punishment and a harm label generated by the identical two branches, because the tables themselves share members. A separate clash check adds its own line: 30.1% of charts get both a clash and a friction note, 7.9% get a clash with no friction match at all (like the chart above that once triggered a real bug), and 17.0% get neither.