AMS English / Industry Anatomy

How Gacha Odds and Claw-Machine Payouts Work

Korea's probability-disclosure timeline, the Nexon fine, and a calculator for the economics behind claw-machine prizes.

KOREA IN 60 SECONDS

What changed—and what did not

2015 was self-regulation. Korea's statutory probability-disclosure duty took effect on March 22, 2024. Calling both phases “the 2015 law” hides the difference between an industry promise and a legal duty.

The fine has two parts. Korea's competition authority imposed KRW 11.593 billion over MapleStory and KRW 49 million over Bubble Fighter: KRW 11.642 billion in total. Nexon's court challenge was still pending at this page's July 2026 snapshot.

Disclosure is not an audit. The harder question is whether versioned rate tables, change logs and server draw records let a regulator reproduce what users actually faced.

0.5% × 100 pulls ≠ 50%With independent fixed odds, the chance of at least one success is 1 − (1 − 0.005)100 = 39.42%. The Korean analysis includes an adjustable calculator.
Use the English odds calculator ↓Open the full Korean source file →
TRY THE FIXED-ODDS MODEL

How many pulls make success “likely”?

This models independent pulls with a fixed published rate. It does not model pity, changing rates, guarantees, exchange shops or character-specific pools.

At least one success39.42%
Expected successes0.50
Spend for these pulls300
50% chance139 pulls
90% chance460 pulls
95% chance598 pulls

Expected value is an average across many repeated experiments. It is not a guarantee that this account will receive that many items.

THE SHORT ANSWER

Is a claw machine skill, luck—or a programmed payout?

All three exist. Japanese arcades commonly mix skill-based two-prong setups with machines whose manuals allow separate weak/strong grip settings and payout management. A near-win alone cannot prove manipulation; repeated observations can reveal the operating pattern.

WHAT THE MACHINE CAN CONTROL
GripWeak and strong claw voltage/strength can be configured on documented models.
CarryHolding strength during lift or horizontal return may differ from the initial grab.
Payout cycleSome control boards support strong-grip/random modes and counters linked to wins.
GeometryBar spacing, rings, prize orientation and staff repositioning can make a game genuinely skill-based—or effectively unwinnable.
JAPAN: TWO DIFFERENT GAMES UNDER ONE NAME
Skill / positioning setup

Often two prongs. Progress remains after each move; the goal is to rotate, slide or tip the prize. Staff assistance is part of the arcade culture.

Grip / payout-managed setup

Common in large three-prong plush machines. The grab can look convincing, then weaken during lift or travel. The average cost per win matters more than a single attempt.

Important: “common” is not “universal.” Machine model, operator settings and prize layout decide the real game.

STORE ECONOMICS CALCULATOR

Change the assumptions. “Attempts per prize” is an observed average—not a guide to hiding manipulation.

Monthly plays
4,050회
Revenue
₩4,050,000
Prize cost
₩1,462,500
Scenario surplus
₩-2,912,500

Fixed costs exceed revenue even before prize costs.

This is not accounting profit. It subtracts only the prize and fixed costs you entered; card fees, taxes, repairs, staffing and financing are excluded unless included in fixed costs.

HOW A CUSTOMER CAN AUDIT THE GAME
  1. Watch several completed plays before paying. Does grip consistently weaken at the same phase?
  2. Prefer games where each move leaves visible progress; avoid resets that erase all progress.
  3. Set a cash limit first. Near-wins are not evidence that the next play will win.
  4. Ask staff for repositioning or a demonstration. Refusal is information, too.
  5. Record model, shop, time, attempts and outcome before making a manipulation claim.
SOURCES & LIMITS

Machine manuals prove that controls exist; they do not prove how a specific shop configured a specific machine. Japanese industry rules also govern prize value and compliant prize distribution.

한국어 전체 분석 보기 →