Formal Specification
Red and White Rose — Formal Specification
Sweden · chase-tag · Confidence: high
← Back to game pageImplementation Suitability
Suitability: adaptation-requiredFidelity: discretizedGap-fills: 5
# Röda vita rosen — AI-optimized specification
## General idea
A traditional Swedish scouting, deduction, and territory espionage game based on Astrid Lindgren's *Kalle Blomkvist* lore. One team ("The Red Rose") hides a secret physical treasure and attempts to evade capture while feeding deceptive intelligence; the opposing team ("The White Rose") patrols the environment, captures opposing agents, and interrogates them under the strict "Rule of Deception" (exactly one true clue and two false clues per interrogation) to deduce the treasure's exact coordinate before time expires.
## Formal model
- **Abstraction taken:** Open town/forest terrain is discretized into an undirected planar graph consisting of named location zones [gap-fill]. Hiding the treasure is modeled as assigning a hidden state variable . Interrogation is formalized as an epistemic query where the captured agent generates 3 logical propositions over , constrained so that exactly 1 is true and 2 are false.
- **Player/agent count:** 2 teams of 4 agents (Team A and Team B; 8 agents total).
- **Turn model:** Phase 1 (Hiding) is 1 setup step; Phase 2 (Search & Interrogation) operates across synchronous discrete time ticks ().
## Formal rules
R1. **Setup:** The map is a graph with zones. Each zone has static attributes (e.g. `region`: `{NORTH, SOUTH, EAST, WEST}`, `terrain`: `{WOODS, RUINS, WATER, MEADOW}`). Team A is designated `HIDING_TEAM`; Team B is `SEARCHING_TEAM`.
R2. **Phase 1 (Hiding the Treasure):**
- Team A selects a hiding zone . The value is added to the private information set of all Team A agents.
- Team A agents disperse across .
R3. **Phase 2 (Patrol & Movement):**
- On each tick , all agents simultaneously move along edges to adjacent vertices or hold position.
- Searching agents patrol to find hiding agents or search vertices for the treasure.
R4. **Capture & Immunity:**
- If at the end of tick , a searching agent and an active hiding agent occupy the same vertex , the hiding agent is marked `CAPTURED`.
- If multiple searchers or hiders occupy , 1 hider is captured [gap-fill].
R5. **The Interrogation (*Förhör*):**
- The captor interrogates the captured prisoner. The engine prompts the prisoner (or automated generator) to output 3 distinct location propositions regarding (e.g., ; ; ).
- Rule of Deception: Exactly 1 proposition must evaluate to TRUE on , and exactly 2 propositions must evaluate to FALSE [gap-fill].
- All 3 propositions are delivered to the searching team.
- Release: The prisoner receives an immunity window of ticks [gap-fill] during which they cannot be captured, and must move away.
R6. **Treasure Search Action:**
- Any searching agent at vertex may execute `inspect_zone()`.
- If , the treasure is discovered immediately, and the Searching Team wins the round.
R7. **Round Expiration & Role Swap:**
- If ticks elapse without the treasure being discovered [gap-fill], the Hiding Team wins the round.
- In Round 2, teams swap roles. The team that finds the treasure in fewer ticks (or survives longest while hiding) wins the match.
## Edge cases
- Tie at the end condition: If both teams find the treasure in the exact same number of ticks during their respective search rounds, the match is a draw.
- A player with no legal move: Graph is connected; agents can always move to an adjacent vertex or wait.
- Repetition / cycles: The search phase is strictly bounded by the monotonic tick limit .
- Simultaneous claims on the same resource or position: Multiple agents can occupy the same zone; capture priorities resolve atomically at the end of each tick.
- Player elimination mid-round: Captured players are never permanently eliminated; they are interrogated, granted 3 ticks of immunity, and rejoin the evasion field.
- Resource exhaustion (pieces, tokens, space, time): Only 1 treasure exists; time bounded by .
- Termination guarantee: Enforced by the finite tick ceiling per round.
## State representation
- **Persistent state:**
- `round`: 1 or 2.
- `team_roles`: `{Team_A: HIDING, Team_B: SEARCHING}`.
- `round_1_ticks_to_win`: optional integer.
- **Per-turn ephemeral state:**
- `tick`: integer .
- `treasure_location`: (hidden from searching team).
- `agent_positions`: map of agent ID .
- `immunity_timers`: map of agent ID .
- `clue_log`: list of revealed proposition triplets.
- **Observable information:**
- Hiding team observes: , positions of all teammates, and positions of searching agents in the same or adjacent zones.
- Searching team observes: positions of all teammates, positions of visible hiders in the same zone, all revealed clue triplets, and visited zones.
## Action space
- `move(target_zone)`: step to adjacent graph vertex.
- `submit_clues(c1, c2, c3)`: executed by captured hider (constrained by 1-True / 2-False rule).
- `inspect_zone()`: search current vertex for .
## Playability notes
- Single-player-vs-AI: Highly engaging — player as Master Detective analyzing clue consistency using propositional constraint solvers (SAT / Bayesian inference) to pinpoint the treasure while commanding searchers.
- Group games (n>4): The 4v4 format reduces to 1 Searcher vs 1 Hider on a smaller 8-zone graph while preserving the full deductive clue-elimination mechanic.
- Physical-skill games: Physical neighborhood sprinting and stealth concealment are abstracted into graph navigation and line-of-sight zone detection.
- What is lost in the adaptation: The childhood literary roleplay of Astrid Lindgren's small-town alleys, handwritten parchment notes, secret passwords, and physical hideouts.
## Implementation hints
Moderate complexity. The deductive clue generator and solver can be implemented cleanly using simple propositional filtering over the 16 candidate vertices. First milestone: build the 16-node graph, implement the 1-True/2-False clue validator, and verify that 3 distinct interrogations reduce the candidate search space of to zones.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Red and White Rose (Sweden). (n.d.). Games of Countries. https://gamesofcountries.com/games/se-sweden/roda-vita-rosen
@misc{gamesofcountries-se-sweden-roda-vita-rosen,
title = {Red and White Rose (Sweden)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/se-sweden/roda-vita-rosen}
}