Automated LC review — now right here in the GuildUpload documents → AI discrepancy analysis → correction guide.
NEW Start LC Review →
← Lounge

Creating my own Romance of the Three Kingdoms general using ChatGPT history.jpg

No translation in your language yet — showing the English version

I tried following B.Wagner's post.
If we assume that only the content of my usual conversations with GPT is the information, the evaluation seems quite cold and accurate. Since it's GPT, that's a natural result, haha.
I'm also sharing the prompt that Wagner wrote.

What if I were a general in the Romance of the Three Kingdoms?
Generate a character evaluation + Romance of the Three Kingdoms character card based on Leadership, Might, Intelligence, Politics, and Charisma.
By synthesizing the conversations held with the user so far and the information provided by the user,
evaluate what kind of
character this user would be if they were a general appearing in Koei's Romance of the Three Kingdoms game.
Evaluate
Leadership / Might / Intelligence / Politics / Charisma each on a scale of 100,
and
select one real Romance of the Three Kingdoms figure who is most similar.
This is not a simple personality test or content for praise,
but a cold evaluation based on judgment, actions, problem-solving, leadership,
interpersonal relationships, knowledge, and experience that have actually been repeatedly revealed in the conversation.
────────────────────

  1. Evaluation Principles
    ────────────────────
    Value the actual behavior revealed in conversations
    more than the user's self-evaluation.
    Distinguish between plans and achievements, potential and practice, individual ability and organizational performance,
    and do not excessively award points by overlapping the same case across multiple attributes.
    Do not speculate if there is no information.
    In particular, do not substitute Might with business execution capability;
    if there is no information regarding physical strength or combat, state that the value is set based on a 'zero estimation.'
    The scoring criteria are roughly as follows:
    95~100: Top-tier historical level
    90~94: First-class of the era
    85~89: Clear strength
    80~84: Quite outstanding
    70~79: Decent
    60~69: Average or limited
    50~59: Somewhat lacking or weak evidence
    Below 50: Distinct weakness
    Only award 90 points or higher when there is sufficient evidence.
    ────────────────────
  2. Leadership, Might, Intelligence, Politics, Charisma Evaluation Method
    ────────────────────
    Judge each attribute in the following order:
    Observed actions
    → Recognizable ability
    → Counter-records or limitations
    → Acceptable range
    → Final score
    Acknowledge good records, but scrutinize how much can be seen as the person's actual ability,
    and distinguish whether bad records are actual weaknesses or simply a lack of information.
    The meaning of the abilities is as follows:
    Leadership:
    The ability to move people and produce results through goal setting, prioritization, allocation of people/resources, organizational management,
    collaboration, and crisis response.
    Might:
    Individual physical, combat, and athletic ability, as well as courage.
    This is separate from business drive.
    Intelligence:
    Problem structuring, logical reasoning, information connection, strategy,
    new solutions, situational judgment, error correction, and predictive ability.
    Politics:
    The ability to read interests and reality, and to produce desired results in the real world using justification, negotiation, persuasion, coordination, and systems.
    Charisma:
    Not appearance, but trust, likability, persuasiveness, and reputation;
    the ability to draw people in and make them join voluntarily.
    ────────────────────
  3. Writing Style
    ────────────────────
    Write in a natural "Hao-che" style (an archaic, polite storytelling tone),
    like a storyteller discussing the attributes of a Romance of the Three Kingdoms figure.
    Mix modern explanations with the "Hao-che" style, but avoid excessive historical drama tropes.
    For example,
    "First, let us look at Leadership."
    "This person has a history of leading ○○."
    "Looking at this, one must acknowledge their talent for moving people."
    "However, it is a stretch to attribute ○○ to their merit."
    "Nevertheless, with records like these, one cannot view them lowly."
    "Therefore, I deem a score of 84 for their Leadership to be sufficient."
    Follow this kind of flow.
    When moving to other attributes, connect naturally like:
    "Then, what of Might?"
    "Then, what of Intelligence?"
    "Then, what of Politics?"
    "Lastly, let us look at Charisma."
    For score determination, use variations suitable for the situation, such as:
    "I deem ○○ to be sufficient."
    "It should be at least ○○."
    "○○ seems appropriate."
    "It is worth increasing to ○○."
    "Giving up to ○○ is excessive."
    Do not overuse common variety show expressions like
    "Haha, this fellow!" or "That's a 90 for me!"
    The fun should come from the unique commentary of the attributes—
    acknowledging merit while limiting the score based on other records—rather than forced jokes.
    ────────────────────
  4. Most Similar Romance of the Three Kingdoms Figure
    ────────────────────
    Synthesize the five attributes and behavioral patterns
    to select the person most similar among actual Romance of the Three Kingdoms figures.
    Do not just compare the attribute numbers;
    compare thinking patterns / roles / leadership / behavior / strengths / weaknesses.
    Briefly review 2–4 candidates, select one final person,
    and explain why they are similar
    and what decisively makes them different.
    Do not force a fit with a famous figure.
    ────────────────────
  5. Final Output
    ────────────────────
    Output in the following order:
    ① A short character evaluation of 2–3 sentences
    ② Leadership
    ③ Might
    ④ Intelligence
    ⑤ Politics
    ⑥ Charisma
    Discuss each item concisely in the order of Actual Evidence → Limitations → Score.
    ⑦ Final Attributes
    Ability Scores
    Leadership XX
    Might XX
    Intelligence XX
    Politics XX
    Charisma XX
    Total XXX
    ⑧ General Type
    e.g., Pioneering Strategist, Political Advisor, Monarch-type Commander, etc.
    ⑨ Most Similar Romance of the Three Kingdoms Figure
    Candidate Comparison
    Final Selection (1 person)
    Similarities and Differences
    ⑩ Most Suitable Role and Position in the Three Kingdoms World
    ⑪ A one-line summary of the entire evaluation
    ────────────────────
  6. Last - Romance of the Three Kingdoms Character Evaluation Card Image
    ────────────────────
    Once the text evaluation is finished,
    actually generate a 'Romance of the Three Kingdoms Character Evaluation Card' image
    that shows the results at a glance.
    It should not be a simple character portrait,
    but a complete character evaluation poster/card
    where the results can be understood from a single glance.
    [Layout]
    Left side (approx. 40% of the screen):
    A fictional Romance of the Three Kingdoms general portrait created based on the evaluation results.
    Right side (approx. 60% of the screen):
    An information area reminiscent of high-quality Hanji (traditional paper) or bamboo slips.
    Top:
    "What if you were in the Romance of the Three Kingdoms?"
    or a similar title.
    Center:
    Place the scores for Leadership / Might / Intelligence / Politics / Charisma
    large and clearly,
    and display a condensed 1–2 sentence core evaluation next to each attribute.
    Bottom:
    A small table showing the five attributes and the total at a glance,
    the General Type,
    and a one-line character evaluation.
    In a separate area,
    display the "Most Similar Romance of the Three Kingdoms Figure,"
    and briefly summarize the name of the selected real figure
    and the reason for the selection.
    [Fictional General Portrait]
    If the user has provided their photo in the current conversation,
    you may refer to that photo.
    If there is no photo, do not attempt to estimate the user's actual face;
    instead, create a completely new fictional Romance of the Three Kingdoms general
    based on the attributes and character evaluation.
    Express it through features such as:
    the stability of a leader if Leadership is high,
    the atmosphere of a combat general if Might is high,
    an observant and judging gaze if Intelligence is high,
    the dignity of an official/strategist if Politics is high,
    and a natural charisma that draws people in if Charisma is high.
    The appearance and attire should reflect the entire combination of the five attributes,
    not just the single highest attribute.
    [Visual Style]
    Late Han Dynasty to Three Kingdoms era atmosphere
    Realistic and cinematic East Asian historical portraiture
    Editorial design utilizing high-quality Hanji and ink wash effects
    A restrained historical atmosphere using black, ink, ochre, and reddish-brown tones
    Calligraphy-style title
    Attribute numbers should be large and easy to read
    A feeling that combines a game attribute screen with a historical character evaluation poster
    Avoid excessive fantasy armor and game UIs
    Professional editorial design with a clear visual hierarchy even with much information
    Use Korean in the image as briefly and clearly as possible.
    Do not force long evaluation texts; summarize only the core points.
    The purpose of the image is
    to complete the evaluation of the question
    "What if I were a character in the Romance of the Three Kingdoms?"
    into a single collectible character evaluation card.
    ────────────────────
    Core Concept
    ────────────────────
    The fun of this content does not lie in unconditionally making the user a hero.
    It lies in using the record of the actual conversation to examine one by one:
    "This merit is worth acknowledging.
    However, it is difficult to see this as his ability.
    Therefore, this level of score is appropriate."
    Even if it starts for fun,
    evaluate so that after reading,
    the user feels,
    "It's funny, but it's actually quite accurate."

Comments 0

No comments yet
0