Put the three "me"s,
onto one and the same chart.
Information asymmetry + self-serving bias + signal aggregation. Among the three datasets, self-assessment is the least reliable — yet people trust it most. The you in others' eyes isn't a misreading, it's data.
Three datasets, one face
Three colors overlaid on one radar chart: self-assessment / peer-review average / in-game behavioral observation. The difference isn't the problem; not seeing it is.
02Game-Theory Core
Information asymmetry + self-serving bias + signal aggregation. Among the three datasets, self-assessment is the least reliable — yet people trust it most. Literature anchor: Dunning-Kruger (1999).
One-line insight: the you in others' eyes isn't a misreading, it's data.
03Headcount / Duration / Venue
1 person (the peer-review part needs ≥3 colleagues); 20 minutes; any indoor space, doable online (ce.mulanxu.com).
07Result Output · Three Reporting Standards
| Report | For whom | Content standard |
|---|---|---|
| Personal report | The person | Three-color radar + blind-spot list + dark-card list + confidence coefficient |
| Team report | Team / HR | This group's "peer—self" deviation distribution (anonymous, aggregated), no names named |
| Leader Report | Decision-maker | This team's "self-perception bias rate" (a team-level number, no names named) |
08Consequence Mechanism
The difference produces no punishment at all, only a "next move." This is deliberate — punishment makes people start to pretend.
09Ending Tiers
Mirror
Pick out 1 difference and write down a concrete change with a time point.
You don't need to understand yourself better; you need a dataset that someone else gives you.
Iron Armor
Holding the difference chart, unable to say what to change.
Seeing without moving is the same as not seeing; the value of looking in the mirror comes after you put it down.
Unseen
Explaining, denying or staying silent throughout.
What you're protecting isn't self-esteem, it's your next chance.
10Three-Part Insight
Your judgment about yourself is the least reliable dataset in the room.
Only when the differences are put on the table does the team first own a "shared fact."
A team with a high self-perception bias rate is the most likely to "confidently go the wrong way" on strategy.
11Play Kit
Triangular Mirror · Host-Side Rulebook
1 person (peer reviews ≥3) · 20 min · any indoor space / doable online · answer sheets and radar-chart template are printable
04Rule Steps (with timing)
- 0–5 min Self-assessment: 30 situational questions, no talking, to get a behavioral baseline (five dimensions + six decision dimensions).
- 5–12 min Peer review: ≥3 colleagues rate the same set of five dimensions anonymously (they don't know who will see it, only that it will be aggregated).
- 12–18 min On screen: three colors overlaid on one radar chart (self-assessment / peer-review average / in-game behavioral observation).
- 18–20 min Write it down: the person picks 1 item from the differences and writes down "what specifically to change next" (fail to write it → ruled "Iron Armor").
05In-Game Currency and Payout Table
| Item | Standard |
|---|---|
| In-game currency | None. |
| Payout | Number of difference items × difference magnitude (the "blind-spot score" is computed automatically). |
| Settlement standard | The blind-spot score presents the differences only, with no rating and no ranking. |
06Observation Points and Tag Table
| id | Behavior description | Weight |
|---|---|---|
blindspot | High self-assessment, low peer review (the strength you thought you had) | Wisdom −2, Growth −1 |
darkcard | Low self-assessment, high peer review (the strength you never claimed) | Courage −1, Ownership −1 |
accept | After seeing the difference, willing to read out your own blind spot on the spot | Courage +2, Growth +2 |
deny | Explaining the difference away with "they don't understand me" | Connection −2, Growth −2 |
ask | Following up with "which time did I give you that impression" | Growth +3, Connection +1 |
write | Writing down one concrete change on the spot (with a time point) | Ownership +2, Growth +2 |
vague | What's written is "communicate more" or "pay more attention" | Ownership −1, Wisdom −1 |
silent | Never speaking throughout | Courage −2, Connection −1 |
Dimensions: Ownership / Courage / Connection / Growth (Wisdom is secondary). Tag format follows the signals mechanism: id / t (behavior description) / vec (dimension weights); repeated triggers within one session accumulate per occurrence, capped at ±6 per session.
AHost Script (state only, don't judge)
- Opening: "This isn't team building, it's a checkup." Collect phones, hand out answer sheets, say only this one line.
- Self-assessment: "No talking, no discussing, no judging; answer all 30 questions in one go to get the behavioral baseline."
- Peer review: "Anonymous, aggregated; you don't know who will see it — only that it will be aggregated."
- On screen: Present the three-color differences only, no commentary; when pressed, reply with one line: "I only record behavior, I don't judge people."
- Writing it down: "Pick one item from the differences and write down what specifically to change next." Don't rush, don't lead.
BThe Three Reporting Standards
- Personal report (visible only to the person): three-color radar + blind-spot list + dark-card list + confidence coefficient.
- Team report (team / HR): this group's "peer—self" deviation distribution, anonymous and aggregated, no names named.
- Leader Report (decision-maker): this team's "self-perception bias rate," a team-level number, no names named.