2nd/10_Wiki/Topics/AI_and_ML/제로잉 (Getting Zero-ed).md

---
id: wiki-2026-0508-제로잉-getting-zero-ed
title: 제로잉 (Getting Zero-ed)
category: 10_Wiki/Topics
status: verified
canonical_id: self
aliases: [Getting Zero-ed, Zero-ed Out, 제로잉, 영점잡기]
duplicate_of: none
source_trust_level: A
confidence_score: 0.85
verification_status: applied
tags: [game-design, rts, balance, attrition]
raw_sources: []
last_reinforced: 2026-05-10
github_commit: pending
tech_stack:
  language: typescript
  framework: ECS/Telemetry
---

# 제로잉 (Getting Zero-ed)

## 매 한 줄
> **"매 player 의 economy/army 가 0 으로 수렴하는 collapse state"**. RTS / wargame 의 매 attrition end-state 로, recovery 가 mathematical 으로 불가능한 지점. 매 design 관점에서 매 surrender threshold + comeback mechanic 의 balance 문제. WARNO, StarCraft, Supreme Commander 의 매 핵심 lose-condition.

## 매 핵심

### 매 zero-ed 의 정의
- **Economic zero**: 매 income < upkeep, reserve = 0 — 매 next unit produce 불가
- **Army zero**: 매 combat force = 0, defenseless — 매 base raze 매 frame
- **Recovery threshold**: 매 economy 가 minimum production 회복 비용 미만
- **Mathematical lock**: 매 enemy income > my max income — recovery 불가능

### 매 collapse spiral
1. Decisive engagement loss → army -50%
2. Map control 상실 → income -30%
3. Production 의 idle → tech lag
4. Counter-attack 막을 force 부족 → base loss
5. 매 income < 0 → zero-ed

### 매 응용
1. Surrender prompt timing — 매 mathematical lock 감지 시 GG 권장.
2. AI difficulty — 매 plateau 직전 rubber-band.
3. Replay analytics — 매 zero-ed timestamp 가 match 의 climax.

## 💻 패턴

### Zero-ed detection
```typescript
type PlayerState = {
  economy: { income: number; upkeep: number; reserve: number }
  army: { combatPower: number; production: number }
  map: { controlPct: number }
}

function isZeroed(p: PlayerState, enemy: PlayerState): boolean {
  const netIncome = p.economy.income - p.economy.upkeep
  const cantBuild = p.economy.reserve < MIN_UNIT_COST && netIncome <= 0
  const noArmy = p.army.combatPower < enemy.army.combatPower * 0.1
  const lockedOut = p.economy.income < enemy.army.combatPower * UPKEEP_PER_POWER
  return cantBuild && noArmy && lockedOut
}
```

### Recovery threshold calc
```typescript
function recoveryThreshold(p: PlayerState, ctx: GameCtx): number {
  // 매 economy 회복에 필요한 minimum reserve
  const baseTickCost = ctx.baseProductionCost
  const ramp = ctx.tickRampMultiplier
  // 매 5 ticks 안에 net positive 회복 가능한가
  const need = baseTickCost * 5 - p.economy.income * 5
  return Math.max(0, need)
}

function canRecover(p: PlayerState, ctx: GameCtx): boolean {
  return p.economy.reserve >= recoveryThreshold(p, ctx)
}
```

### Surrender prompt logic
```typescript
class SurrenderPrompt {
  private confirmedZeroedAt: number | null = null

  check(p: PlayerState, enemy: PlayerState, now: number): boolean {
    if (!isZeroed(p, enemy)) {
      this.confirmedZeroedAt = null
      return false
    }
    if (this.confirmedZeroedAt === null) {
      this.confirmedZeroedAt = now
      return false
    }
    // 매 30 seconds 동안 zero-ed 유지 → prompt
    return now - this.confirmedZeroedAt > 30_000
  }
}
```

### Comeback mechanic
```typescript
function comebackBuff(p: PlayerState, enemy: PlayerState): Buff | null {
  const ratio = p.army.combatPower / (enemy.army.combatPower + 1)
  if (ratio > 0.4) return null
  // 매 underdog buff
  return {
    incomeMultiplier: 1.0 + (1 - ratio) * 0.3,
    productionSpeedMultiplier: 1.0 + (1 - ratio) * 0.2,
    durationTicks: 60,
  }
}
```

### Map control to income
```typescript
function calcIncome(controlled: Region[], baseIncome: number): number {
  const total = controlled.reduce((s, r) => s + r.value, 0)
  // 매 diminishing returns
  return baseIncome + Math.sqrt(total) * 12
}
```

### Spiral telemetry
```typescript
function logSpiralEvent(p: PlayerState, prevTick: PlayerState, tick: number) {
  const armyDelta = p.army.combatPower - prevTick.army.combatPower
  const econDelta = p.economy.income - prevTick.economy.income
  if (armyDelta < -prevTick.army.combatPower * 0.3) {
    telemetry.emit("decisive_loss", { tick, lossPct: armyDelta / prevTick.army.combatPower })
  }
  if (econDelta < -prevTick.economy.income * 0.2) {
    telemetry.emit("economy_collapse", { tick, dropPct: econDelta / prevTick.economy.income })
  }
}
```

## 매 결정 기준
| 게임 mode | Zero-ed handling |
|---|---|
| Ranked 1v1 | 매 strict zero detection + GG prompt @ 30s lock |
| Casual | 매 soft comeback buff at <40% army ratio |
| Co-op vs AI | 매 ally rescue mechanic, no auto-surrender |
| Tournament | 매 surrender 자유, no auto — strategic tilt |
| Tutorial / Campaign | 매 mission-fail trigger, retry |

**기본값**: detect + comeback buff + GG prompt at 30s confirmed zero.

## 🔗 Graph
- 부모: [[전투 전술(Battle Strategies)]] · [[Game_Design]]
- 변형: [[Permanent_Loss]] · [[Power Creep (Content Treadmills)|Power_Creep]]
- 응용: [[Eugen Systems 모딩 매뉴얼|WARNO]] · [[Eugen Systems 모딩 매뉴얼|Eugen_Systems]] · [[알비온_온라인(Albion_Online)]]
- Adjacent: [[Telemetry_Balancing]] · [[Combat_Timeline_Difficulty_Scaling]]

## 🤖 LLM 활용
**언제**: post-match analysis 의 collapse moment 설명, design tuning rationale 작성.
**언제 X**: 매 frame-tight detection — 매 deterministic check.

## ❌ 안티패턴
- **No surrender prompt**: 매 zero-ed 후 5분 farming — 매 user frustration.
- **Over-tuned comeback**: 매 강한 underdog buff → skill 무의미.
- **Hard recovery cliff**: 매 binary recovery, no gradient → 매 frustration spike.
- **Missing telemetry**: 매 zero-ed timestamp 미수집 — design tuning 불가.

## 🧪 검증 / 중복
- Verified (Eugen WARNO matchmaking design notes, StarCraft 2 GG protocol).
- 신뢰도 B+ (game-specific terminology).

## 🕓 Changelog
| 날짜 | 변경 |
|---|---|
| 2026-05-08 | Phase 1 |
| 2026-05-10 | Manual cleanup — zero-ed collapse detection + comeback patterns |