Formal Specification
Catch the Handkerchief / Handkerchief Tag — Formal Specification
Türkiye · chase-tag · Confidence: low
← Back to game pageImplementation Suitability
Suitability: engine-readyFidelity: discretizedGap-fills: 2
# Mendil Kapmaca — AI-optimized specification
## General idea
A high-speed tactical tag and sprint game discretized onto a 1D coordinate line. Two agents start at opposite ends and run to the center (0), where they engage in a psychological standoff. Once one agent grabs the handkerchief, the game transitions to a chase where the grabber sprints back to their starting line while the other pursues.
## Formal model
- **Abstraction taken:** The rectangular court is abstracted onto a 1D coordinate axis ranging from `-14.0` (Agent 0's baseline) to `+14.0` (Agent 1's baseline). The center coordinate `0.0` represents the handkerchief. [gap-fill] The physical bluffing and reflexes are modeled as a tick-based action selection game.
- **Player/agent count:** 2 active agents per duel (representing the called pair).
- **Turn model:** Real-time ticked. Each tick, both agents simultaneously declare their movement speed and action, which are then resolved.
## Formal rules
R1. State representation:
- `pos_0: float` (Agent 0's position, initialized to `-14.0`).
- `pos_1: float` (Agent 1's position, initialized to `+14.0`).
- `handkerchief_grabbed: boolean` (initialized to false).
- `grabber: integer` (-1, 0, or 1).
- `active_phase: integer` (0: Sprint to Center, 1: Standoff, 2: Chase).
- `time_limit: integer` (standoff limit, initialized to 40 ticks).
R2. Sprint to Center (Phase 0):
- Both agents simultaneously select a move distance.
- Each tick:
- `pos_0 = min(0.0, pos_0 + speed(0))` where `speed` is a function of agent skill and fatigue.
- `pos_1 = max(0.0, pos_1 - speed(1))`.
- Once both `pos_0 == 0.0` and `pos_1 == 0.0`, the game enters Phase 1 (Standoff).
R3. Standoff (Phase 1):
- Both agents stand at `0.0`.
- On each tick, both agents simultaneously select an action from `action_set ∈ {WAIT, FEINT, GRAB}`.
- Resolution table:
- If both choose `WAIT`: nothing happens, `time_limit` decreases by 1.
- If one chooses `WAIT` and one chooses `FEINT`: the feinting player gains a reflex bonus on the next tick.
- If one chooses `GRAB` and the other chooses `WAIT` or `FEINT` (and feinter has no bonus): the grabbing player successfully grabs the handkerchief: `handkerchief_grabbed = true`, `grabber = p`, phase transitions to Phase 2 (Chase).
- If one chooses `GRAB` and the other has a reflex bonus from a previous `FEINT`: the bonused player intercepts the grab: `handkerchief_grabbed = true`, `grabber = opponent(p)`, phase transitions to Phase 2.
- If both choose `GRAB` simultaneously:
- [gap-fill] The grabber is determined stochastically based on respective skills: `P(Agent 0 grabs) = skill(0) / (skill(0) + skill(1))`.
- If `time_limit == 0`:
- A coin toss decides the grabber, and Phase 2 begins.
R4. Chase (Phase 2):
- The grabber `g` attempts to reach their baseline (`-14.0` if `g=0`, `+14.0` if `g=1`).
- The pursuer `p = opponent(g)` attempts to close the distance and tag the grabber.
- On each tick:
- The grabber moves towards their baseline: `pos_g = pos_g + direction_g * speed(g)`.
- The pursuer moves in the same direction: `pos_p = pos_p + direction_g * speed(p)`.
- If at any tick `distance(pos_g, pos_p) <= 0.0` (pursuer overtakes or reaches grabber):
- Pursuer `p` tags the grabber. Pursuer's team wins: `score[p] = 1`, game terminates.
- If the grabber reaches or exceeds their baseline (`|pos_g| >= 14.0`):
- Grabber `g` safely returns. Grabber's team wins: `score[g] = 1`, game terminates.
## Edge cases
- **Tie at the end condition:** Not applicable (first event to trigger wins).
- **A player with no legal move:** Not applicable.
- **Repetition / cycles:** Standoff is strictly bounded by `time_limit` (R3), and the chase phase features players moving strictly towards a baseline, guaranteeing termination.
- **Simultaneous claims on the same resource:** Handled by R3's simultaneous grab check.
- **Player elimination mid-round:** Not applicable.
- **Resource exhaustion:** Not applicable.
- **Termination guarantee:** Guaranteed by the countdown timer in Phase 1 and the monotonic movement to baselines in Phase 2.
## State representation
- Persistent state:
- `pos: Array[2] of float`.
- `handkerchief_grabbed: boolean`.
- `grabber: integer`.
- `phase: integer`.
- `time_limit: integer`.
- Observable information: Fully public, perfect information.
## Action space
- `sprint()`: Phase 0 action.
- `standoff_action(type)`: where `type ∈ {WAIT, FEINT, GRAB}`. Valid only in Phase 1.
- `chase_sprint()`: Phase 2 action.
## Playability notes
- Single-player-vs-AI: Excellent. Can be played against an AI that uses Q-learning or heuristic state values to time feints and grabs.
- Group games: Discretizes to a 1v1 duel which represents the core engine of the match.
- Physical-skill games: Physical reflexes, speed, and acceleration are abstracted into tick-based movements and standoff action matchups.
- What is lost: The physical dodging and twisting of the body during the chase.
## Implementation hints
Low-medium complexity. The transition between the standoff phase (R3) and the chase phase (R4) is the core logic. First milestone: build the 1D movement model and implement Phase 2 (Chase) with a static starting distance, then add the standoff game.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Catch the Handkerchief / Handkerchief Tag (Türkiye). (n.d.). Games of Countries. https://gamesofcountries.com/games/tr-turkiye/mendil-kapmaca
@misc{gamesofcountries-tr-turkiye-mendil-kapmaca,
title = {Catch the Handkerchief / Handkerchief Tag (Türkiye)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/tr-turkiye/mendil-kapmaca}
}