2nd/10_Wiki/Topics/AI_and_ML/게이미피케이션.md

id, title, category, status, canonical_id, aliases, duplicate_of, source_trust_level, confidence_score, verification_status, tags, raw_sources, last_reinforced, github_commit, tech_stack

id	title	category	status	canonical_id	aliases	duplicate_of	source_trust_level	confidence_score	verification_status	tags	raw_sources	last_reinforced	github_commit	tech_stack
wiki-2026-0508-게이미피케이션	게이미피케이션	10_Wiki/Topics	verified	self	Gamification 게임화 Game-like Mechanics	none	A	0.9	applied	game-design behavior motivation ux		2026-05-10	pending	language framework typescript behavioral-design / Octalysis

게이미피케이션

매 한 줄

"매 non-game context에 game-design element를 적용해 user motivation / engagement / behavior를 reshape". 2002 Nick Pelling의 용어 등장 → 2010 Foursquare 배지로 mainstream → 2014 Yu-kai Chou의 Octalysis 8-driver framework → 2026 현재 Duolingo · Strava · Apple Fitness가 industry exemplar이며 LLM-generated personalized quests가 emerging.

매 핵심

매 PBL Triad (Basic)

• Points (P): 매 quantitative score — XP, currency.
• Badges (B): 매 achievement marker — collection / status.
• Leaderboards (L): 매 social comparison — ranked scoreboard.

매 Octalysis 8 Core Drives (Yu-kai Chou)

1. Epic Meaning: 매 큰 mission의 일부 — Wikipedia 기여.
2. Development & Accomplishment: 매 progress / mastery — Duolingo streak.
3. Empowerment of Creativity: 매 player creativity expression — Minecraft.
4. Ownership & Possession: 매 collection / customization.
5. Social Influence: 매 peer comparison / cooperation.
6. Scarcity & Impatience: 매 limited-time / FOMO.
7. Unpredictability: 매 variable reward — gacha, slot.
8. Loss & Avoidance: 매 streak break 회피.

매 응용

1. Duolingo — streak, league, hearts (loss-avoidance).
2. Strava — segment leaderboard, kudos.
3. Apple Fitness — 3 rings + 30-day challenge.
4. Habitica — RPG-style habit tracker.
5. LinkedIn — profile completion %, "All-Star" badge.

💻 패턴

XP / Level System

const xpForLevel = (level: number): number =>
  Math.floor(100 * Math.pow(level, 1.5));

export function levelFromXP(totalXP: number): { level: number; progress: number } {
  let level = 1;
  let acc = 0;
  while (acc + xpForLevel(level) <= totalXP) {
    acc += xpForLevel(level);
    level++;
  }
  const inLevel = totalXP - acc;
  const need = xpForLevel(level);
  return { level, progress: inLevel / need };
}

Streak Tracker

import { differenceInCalendarDays } from 'date-fns';

interface StreakState { current: number; longest: number; lastActiveDate: Date }

export function recordActivity(state: StreakState, now: Date = new Date()): StreakState {
  const gap = differenceInCalendarDays(now, state.lastActiveDate);
  if (gap === 0) return state;                    // already counted today
  if (gap === 1) {                                // consecutive day
    const next = state.current + 1;
    return { current: next, longest: Math.max(next, state.longest), lastActiveDate: now };
  }
  return { current: 1, longest: state.longest, lastActiveDate: now };  // broken
}

Badge Engine (rule-based)

type BadgeRule = { id: string; name: string; check: (s: PlayerStats) => boolean };

const RULES: BadgeRule[] = [
  { id: 'first_blood', name: 'First Win', check: s => s.wins >= 1 },
  { id: 'centurion', name: '100 Wins', check: s => s.wins >= 100 },
  { id: 'streak_30', name: '30-Day Streak', check: s => s.streak >= 30 },
];

export function evaluateBadges(stats: PlayerStats, owned: Set<string>): string[] {
  return RULES.filter(r => !owned.has(r.id) && r.check(stats)).map(r => r.id);
}

Variable Reward (Loot Box) with disclosed odds

type Drop = { item: string; weight: number };

export function rollLoot(table: Drop[], rng: () => number = Math.random): string {
  const total = table.reduce((s, d) => s + d.weight, 0);
  let r = rng() * total;
  for (const d of table) {
    r -= d.weight;
    if (r <= 0) return d.item;
  }
  return table[table.length - 1].item;
}

// 매 disclosed odds (regulatory: KR / CN / BE / NL)
const TABLE: Drop[] = [
  { item: 'common',    weight: 70 },
  { item: 'rare',      weight: 25 },
  { item: 'epic',       weight: 4.5 },
  { item: 'legendary',  weight: 0.5 },
];

Leaderboard (Redis ZSET)

import { Redis } from '@upstash/redis';
const redis = Redis.fromEnv();

export async function recordScore(userId: string, score: number) {
  await redis.zadd('leaderboard:weekly', { score, member: userId });
}

export async function topN(n = 10) {
  return redis.zrange('leaderboard:weekly', 0, n - 1, { rev: true, withScores: true });
}

export async function rankOf(userId: string) {
  return redis.zrevrank('leaderboard:weekly', userId);
}

Personalized Quest (LLM-generated)

import Anthropic from '@anthropic-ai/sdk';
const client = new Anthropic();

export async function dailyQuest(profile: { skills: string[]; weak: string[]; mood: string }) {
  const msg = await client.messages.create({
    model: 'claude-opus-4-7',
    max_tokens: 200,
    system: 'Generate ONE daily quest for a learning app. Output JSON: {title, goal, xp, steps[]}',
    messages: [{ role: 'user', content: JSON.stringify(profile) }],
  });
  return JSON.parse((msg.content[0] as { text: string }).text);
}

매 결정 기준

상황	Approach
Education app	Streak + XP + League (Duolingo)
Fitness	Goal-rings + monthly challenge
Productivity	Habit-RPG (Habitica)
Enterprise training	Badge + leaderboard, opt-in
Health-critical	Avoid scarcity / FOMO — ethical-only

기본값: PBL + 1~2 Octalysis drives matched to context, with disclosed-odds + opt-out.

🔗 Graph

• 부모: Game_Design
• 변형: Self_Determination_Theory
• Adjacent: Operant_Conditioning

🤖 LLM 활용

언제: personalized quest generation, badge naming, encouragement copy, weakness-targeted exercise sequencing. 언제 X: 매 reward-schedule design — RL / behavioral economist domain (LLM은 hidden bias 위험).

❌ 안티패턴

• Pointsification: 매 PBL만 얹기 — intrinsic motivation 없으면 6주 후 fatigue.
• Dark patterns: 매 streak shame, FOMO-pressure, sunk-cost manipulation.
• One-size-fits-all: 매 모든 user에 leaderboard — competitive-averse user 이탈.
• Reward inflation: 매 점차 큰 보상 → 매 동일 효과 위해 escalation 필요.

🧪 검증 / 중복

• Verified (Yu-kai Chou Actionable Gamification, Self-Determination Theory, Duolingo Engineering Blog 2024).
• 신뢰도 A.

🕓 Changelog

날짜	변경
2026-05-08	Phase 1
2026-05-10	Manual cleanup — PBL + Octalysis 8 drives + ethical guardrails 정리