Overview
Updated on March 25, 2026 11:30 PM PST
Korapsyon is a social deduction game for 9–16 players set in a fictional Philippine government. Players hold hidden roles — good politicians trying to build public projects, or corrupt politicians quietly looting while appearing legitimate.
The Core Loop
The game runs in turns called Years. Each Year has two phases:
- Hidden Phase — secret night actions (stealing, protecting, investigating, auditing)
- Public Phase — governance, open discussion, and a jail vote
If you want the exact moderator script, see Moderator Reference. For the win checks, see Win Conditions.
The core tension is information asymmetry: the corrupt know each other and act in secret; the good must deduce who is corrupt through observable signals — treasury changes, project vote patterns, and COA audit reports.
Design Pillars
Deception through governance Corruption isn't just stealing — it's exploiting public projects for kickbacks while maintaining plausible deniability.
Delayed information Most evidence of corruption is one Year old by the time it becomes public. The COA reveals last Year's audit, not the current one.
Asymmetric pressure Good must complete enough public works to hit the spending threshold. Bad must quietly fill their Corruption Fund before being permanently eliminated.
Stakes escalate over time Players who are jailed twice are permanently eliminated. As the game progresses, the table shrinks, committees get smaller, and every vote carries more weight.
Player Count and Targets
| Players | Public Spending Target (Good) | Corruption Fund Target (Bad) |
|---|---|---|
| 9–10 | 16 | 14 |
| 11–12 | 20 | 17 |
| 13–14 | 24 | 20 |
| 15–16 | 28 | 23 |
Base Composition (13 players)
| Role | Count |
|---|---|
| Good Politician | 2 |
| Corrupt Politician | 2 |
| Lawyer | 1 |
| Ombudsperson | 1 |
| COA | 1 |
| Citizen | 6 |
Term Lengths
President term = 2 Years · First jail = 3-year suspension · Second jail = permanent elimination
Changes Summary with Timeline
Comparing original single-page handout to current ruleset (March 25, 2026)
Steal Amounts
- Old: Variable range (2 free = 3–5; 1 free = 1–3)
- New: Fixed — 2 free = 4, 1 free = 2
Suppress Audit (new mechanic)
- Old: Didn't exist; COA always got the real audit
- New: Corrupt may suppress during any steal year; costs 1 from steal; COA gets falsified report with blank votes
Witness (new mechanic)
- Old: Ombudsperson had no public action
- New: Once per game, Ombudsperson reveals their result publicly and triggers a runoff jail vote; exposes their identity
Jail System
- Old: Temporary jail only (3-year suspension); no strikes, no permanent elimination
- New: Two-strike system — first jail = 3-year suspension; second jail = permanent elimination with role reveal
COA Audit
- Old: Official spending + kickback yes/no only
- New: Also includes full per-player committee vote breakdown; blanked out when suppressed
Win Conditions
- Old: Bad wins via treasury collapse (Kaban ng Bayan = 0); optional resilience win; Good had a treasury-target win
- New: Treasury collapse and resilience wins removed; Bad wins via private Corruption Fund threshold only
Project Cards
- Old: Separate cost and value fields (could differ)
- New: Cost = value; one number only
Revenue / Tax
- Old: Called "taxes"; suspended players not explicitly addressed
- New: Both "revenue" and "tax" are used; suspended players explicitly contribute nothing
Corruption Fund Tracker (new tracker)
- Old: Didn't exist as a separate tracker
- New: Private moderator tracker — running total of all theft + kickbacks; Bad's sole threshold win condition
First Steal Timing
- Old: Corrupt recognition and first steal were one combined opening step at setup
- New: Setup is identification only (eyes open, see each other, eyes close); stealing happens during the Year 1 Hidden Phase like every other Year
Ombudsperson Investigation Result
- Old: Unspecified what was learned
- New: Alignment only (Good/Bad) — not the specific role; result is private and expires end of Year