It isn't playing games,
it's turning choices into evidence.
Twelve truly playable on-site games, four session packages. Every choice lands as a behavior tag; tags carry dimension weights; the weights converge into three reports — personality, team and decisions. The "result / personality type / consequence / ending / insight" you asked for is all derived from one and the same causal chain, never written by hand.
The whole system runs on a single causal chain
Every game shares one and the same core, with no inventing your own. That is what makes a single session's behavior accumulate, compare and retell — after 151 choices, confidence can reach 95%.
01Causal Chain
choice → behavior tag (signal) → five-dimension weighting → 16 Mulan personality types → ending tiers → insight; a side branch: six decision dimensions + team metrics → team report / Leader Report. The "result / personality type / consequence / ending / insight" the user asks for is all derived from this chain.
02Five Dimensions · Personality
Ownership / Courage / Connection / Wisdom / Growth. Following the established authoritative definitions, unchanged.
03Six Decision Dimensions · Behavior Under Pressure
Information dependency / risk appetite / decision speed / tolerance for dissent / stability under pressure / long-termism.
04Behavior Tag (signal)
Format follows the established signals mechanism: id (the tag's English name) / t (behavior description, for the host's verbatim record) / vec (dimension weights). A game usually defines 8 tags, weight range −3 ~ +3; the same tag triggered repeatedly within one session accumulates per occurrence, capped at ±6 per session.
05Triangulation · Three Perspectives Crossed
In-game behavioral observation × pre-game self-perception (30-question silent baseline) × anonymous peer review (≥3 people). When all three agree → consensus, written into the team report; when they disagree → blind spot, written into the personal report. The degree of agreement converts into that report's confidence coefficient.
06Dual-Track Settlement
| Track | Content | Release Timing |
|---|---|---|
| Track One · Visible | Provisions / copper / market value / inventory / win-loss ranking | Real-time in-game + announced first at settlement |
| Track Two · Invisible | Behavior tags → changes in the five dimensions → decision profile (six-dimension radar) + pressure-state delta | 10 minutes after settlement |
07Ending Tiers · Unified Three Levels
| Level | Name | Determination | Volume |
|---|---|---|---|
| Upper | Mirror | Saw, on the spot, the self that others see, and changed one concrete behavior / wrote down one concrete change there and then | Volume III · Mirror |
| Middle | Iron Armor | Received the difference data but converted it into no action | Volume II · Iron Armor |
| Lower | Unseen | Only won or lost inside the game, taking away no information about oneself | — |
Every ending must carry three insights: for the individual / for the team / for the decision-maker (leader). This is where "an ending that gives everyone an insight" lands.
08Anti-Faking and Rule Violations
- Triple anti-faking detection: social-desirability items / response time / straight-line answering; once detected, weight tips toward the behavioral items, and the confidence is lowered accordingly and stated openly.
- In-session violations: breaking an alliance, swapping cards privately, impersonation → not elimination, but recorded as behavior tags (a broken promise is itself the best data).
- No right-or-wrong principle: the host states behavior only, does not rate it, does not judge personality.
Three tiers, answering three different questions
Tier A hands you a mirror; Tier B lets the mirrors reflect one another; Tier C puts a whole group into real pressure and reflects it. The three tiers can be bought separately or as a bundle.
See Yourself
1 person (peer reviews need ≥3) · 20–45 min. Triangular Mirror, Blind-Spot Auction, Hundred Costs, Compound Gate — four roads all lead to the same face.
View the Tier A guide →See Each Other
6–40 people · 20–40 min. Fault Lines, One Cauldron, Wordless Pact, Complementary Puzzle, Pressure Drill — put relationships into the rules and see through them.
View the Tier B guide →See the Leadership
2–36 people · 45–180 min. Shunxi, Make Camp, Market · Twelve Stalls — low-frequency upheaval, sequential trade-offs, high-frequency feedback, three kinds of pressure each seen through once.
View the Tier C guide →Every game has its own page and rulebook
Open any game: the top screen is the introduction and the game-theory core, the bottom screen is a directly printable host-side rulebook — with timing, scoring, scripts and a materials list.
Triangular Mirror
Stack three datasets — "the me I think I am," "the me others see," "the me I actually am" — onto one and the same chart.
View rulebook→Blind-Spot Auction
Bid with real auction currency on the matter of "how well I know myself" — and see whether you overpay.
View rulebook→Hundred Costs
Both options in every question carry a cost — this is measurement, not a questionnaire.
View rulebook→Compound Gate
Twelve gates, each asking only one question — take it now, or wait a while longer.
View rulebook→Fault Lines
Lay a team's real cognitive disagreements on the table there and then — and see exactly how they get flattened.
View rulebook→One Cauldron
Turn "will the team actually chip in money and effort together" into arithmetic — and arrive at one key number: 62.5%.
View rulebook→Wordless Pact
Not a word written, not a signature given — see how much each of you puts out and how much comes back across the whole team.
View rulebook→Complementary Puzzle
Let everyone see — my weak spot is exactly the strong suit of the person beside me.
View rulebook→Pressure Drill
Watch behavior deform under real pressure — the most accurate setting for behavioral observation, and the only segment that makes people sweat.
View rulebook→Shunxi
A conventional sandbox compares business results; the MulanXu sandbox compares decision behavior — we don't look at who wins, we look at how you win and how you lose.
View rulebook→Make Camp
Give you a troop and a stretch of terrain — how you deploy them is the map in your head of "what matters."
View rulebook→Market · Twelve Stalls
Give you a shop and twelve months. The most honest thing about business is this: every month it tells you once that you were wrong.
View rulebook→Four packages, from 60 minutes to two days
On a tight budget, start with the Icebreaker Package; mid-level teams take the Collision Package; the decision layer takes the Leadership Checkup Package; annual-scale engagements take the Restructuring Package.
Icebreaker Package
60 minWordplay · Odd One Out (warm-up) → Triangular Mirror quick version → Fault Lines, 3 questions → One-Sentence Commitment
Collision Package
180 minSilent Baseline → Triangular Mirror → Fault Lines → One Cauldron → Wordless Pact → Closing
Leadership Checkup Package
180 minSilent Baseline → Complementary Puzzle → Shunxi, 12 rounds → Dual-Track Settlement → Closing
Restructuring Package
2 daysDay 1 full run of Tiers A + B → Day 2 C1 Shunxi + C2 Make Camp + role mapping → pact signing + 90-day debrief checkpoints
Take away a table you can follow directly
Examples below. Each game page has its own game-length timetable; the session-level timetable is for the host to check against on site.
| Time slot | Segment | Host action |
|---|---|---|
| 00–05 | Entry · say only one line: "This isn't team building, it's a checkup." | Collect phones, hand out name badges |
| 05–20 | W Wordplay · Odd One Out (warm-up, 2 rounds) | No commentary; just note "who spoke first" |
| 20–40 | A1 Triangular Mirror quick version (20 self-assessment questions + peer-review summary) | Show differences only, no commentary |
| 40–55 | B1 Fault Lines, 3 questions (question 1 is the anonymous baseline) | Read the question, 30 seconds to answer, publish the distribution |
| 55–60 | Closing Two · One-Sentence Commitment | Each person one line: "Starting today, I give myself permission to ______" |
| Time slot | Segment | Host action |
|---|---|---|
| 00–15 | Entry · Silent Baseline (30 questions, no talking) | Collect phones, hand out answer sheets, say only one line: "Establish the behavioral baseline first" |
| 15–35 | A1 Triangular Mirror (self- and peer-assessment differences on screen) | Show differences only, no commentary |
| 35–60 | B1 Fault Lines, 5 questions | Read the question, 30 seconds to answer, publish the distribution |
| 60–100 | B2 One Cauldron, 6 rounds | Control the pace; switch the public/punishment mechanism at round 4 |
| 100–135 | B3 Wordless Pact, 4 rounds | Rotate roles; publish the return rate from round 3 |
| 135–165 | Team profile on screen + interpretation | Three charts on the big screen + five metrics |
| 165–180 | Closing (pact signing + One-Sentence Commitment) | Print and sign, sync to the AI team space |