2nd/10_Wiki/Topic_General/Game_Design/Combat_Balance_Buff.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-combat-balance-buff	Combat Balance Buff	10_Wiki/Topics	verified	self	combat-buff balance-patch buff-design	none	A	0.88	applied	game-design balance combat patch-design live-ops		2026-05-10	pending	language framework design-doc live-balance

Combat Balance Buff

매 한 줄

"매 underperforming unit / ability 의 매 strength 의 increase". 매 nerf 의 opposite — 매 player frustration 의 minimal 한 buff 가 매 balance team 의 default tool. 매 telemetry-driven (winrate < 47%, pickrate < 5%) buff 의 매 modern live-ops standard.

매 핵심

매 buff 종류

• Stat buff: 매 damage / HP / speed 의 numeric.
• Mechanic buff: 매 cooldown 단축, range 증가.
• Synergy buff: 매 set bonus, faction bonus.
• Quality-of-life: 매 animation 단축, hitbox 확대.

매 결정 기준 (telemetry)

• Winrate: 매 baseline ±3%.
• Pickrate: 매 5-15% target.
• Ban rate (PvP): 매 < 30%.
• Player sentiment: 매 forum / Reddit weighted.

매 응용

1. 매 patch cadence (2-week, monthly).
2. 매 PTR (public test realm) testing.
3. 매 retro-buff 의 long-tail content.

💻 패턴

Buff candidate selector

type UnitTelemetry = {
  unit_id: string;
  winrate: number;       // 0-1
  pickrate: number;      // 0-1
  banrate: number;       // 0-1 (PvP)
  sample_size: number;
};

function bufCandidates(units: UnitTelemetry[]): UnitTelemetry[] {
  return units
    .filter(u => u.sample_size > 1000)
    .filter(u => u.winrate < 0.47 || u.pickrate < 0.05)
    .filter(u => u.banrate < 0.10) // not problem-banned
    .sort((a, b) => a.winrate - b.winrate);
}

Buff magnitude calculator

function suggestBuffPercent(current_winrate: number, target_winrate = 0.5): number {
  const gap = target_winrate - current_winrate;
  // Empirical: ~5% stat buff ≈ 1% winrate shift
  const raw = gap * 5 * 100;
  return Math.round(clamp(raw, 3, 15)); // bounded
}

// Example: winrate 0.42 → suggest +8% damage

Multi-stat buff distribution

function distributeBuff(
  total_buff_budget: number,  // e.g. 10 (%)
  stats: ("damage" | "hp" | "speed" | "cooldown")[]
): Record<string, number> {
  const weights = { damage: 0.4, hp: 0.3, speed: 0.15, cooldown: 0.15 };
  const out: Record<string, number> = {};
  let total_w = stats.reduce((a, s) => a + weights[s], 0);
  for (const s of stats) {
    out[s] = (weights[s] / total_w) * total_buff_budget;
  }
  return out;
}

Patch-note generator

function generatePatchNote(unit: string, changes: Change[]): string {
  const lines = [`## ${unit}`, ""];
  for (const c of changes) {
    const arrow = c.delta > 0 ? "↑" : "↓";
    lines.push(`- ${c.stat}: ${c.before} → ${c.after} (${arrow}${Math.abs(c.delta)}%)`);
  }
  lines.push("", `*Designer note: ${changes[0]?.rationale ?? ""}*`);
  return lines.join("\n");
}

A/B buff validation (PTR)

async function validateBuff(unit_id: string, ptr_data: TelemetryWindow) {
  const before = await getLiveTelemetry(unit_id, "30d");
  const after = ptr_data;

  const winrate_shift = after.winrate - before.winrate;
  const pickrate_shift = after.pickrate - before.pickrate;

  if (winrate_shift > 0.06) return { decision: "scale_back", reason: "overshoot" };
  if (winrate_shift < 0.01) return { decision: "increase", reason: "insufficient" };
  return { decision: "ship", winrate_shift, pickrate_shift };
}

매 결정 기준

상황	Approach
매 winrate 42-47%	+5-8% stat buff
매 pickrate <3%	mechanic / QoL buff
매 niche pick	synergy buff
매 winrate >50% but low pickrate	accessibility buff (animation, range)

기본값: 매 telemetry > 1000 sample + 매 conservative +5% first pass.

🔗 Graph

• 부모: 게임 밸런싱 · LiveOps
• 응용: Defense-Buildings · Damage-Resistance-Platforms
• Adjacent: Evolution-of-the-War-Commander-Combat-Ecosystem

🤖 LLM 활용

언제: 매 patch design, telemetry analysis, magnitude calibration. 언제 X: 매 launch tuning — 매 telemetry 부족.

❌ 안티패턴

• Power creep: 매 매번 buff — 매 이전 unit 의 obsolete.
• Buff knee-jerk: 매 forum 의 single voice 에 react.
• Stack 무시: 매 multiple buff 의 multiplicative 효과.
• Sample 부족: 매 < 1000 의 의미 없음.

🧪 검증 / 중복

• Verified (Riot Games balance team blog, Blizzard patch notes 2024-2025).
• 신뢰도 A.

🕓 Changelog

날짜	변경
2026-05-08	Phase 1
2026-05-10	Manual cleanup — combat balance buff telemetry + magnitude design.