Formal Specification
Khorasan Choobeh Traditional Wrestling — Formal Specification
Iran · wrestling-combat · Confidence: low
← Back to game pageImplementation Suitability
Suitability: adaptation-requiredFidelity: probabilistic-proxyGap-fills: 2
# Koshti ba Choobeh — AI-optimized specification
## General idea
A two-agent traditional grappling contest modeling jacket-and-sash leverage, stamina management, and hip/trip throw techniques. Agents balance grip establishment on the wool jacket against off-balancing momentum to execute instant-victory shoulder throws or accumulate technical points.
## Formal model
- **Abstraction taken:** Grappling is abstracted into discrete 10-second exchange ticks across a 300-second regulation match. Each tick, agents simultaneously choose grip intent and throw/defense tactics, resolved via stochastic functions parameterized by strength, technique, and stamina differential.
- **Player/agent count:** 2 agents (Wrestler 0 and Wrestler 1).
- **Turn model:** Simultaneous tactical commitments per tick.
## Formal rules
R1. Initial Bout State:
- Match duration: `max_ticks = 30` (300 seconds). Overtime adds `18 ticks` (180s).
- Agents start with `stamina = 100`, `grip_level = 0`, `points = 0`, `passivity = 0`.
R2. Legal Action Set per Tick:
- `establish_jacket_grip()`: Seek upper lapel/sleeve control (consumes 8 stamina).
- `establish_sash_grip()`: Seek deep waist belt hold (*Shal*) (consumes 12 stamina; requires `grip_level >= 1`).
- `attempt_throw(technique)`: Execute hip throw (*Pich*), shoulder throw (*Koolandaz*), or leg trip (*Lengeh*). Requires `grip_level >= 1`.
- `defend_sprawl()`: Focus on breaking opponent grips and recovering balance (recovers +6 stamina).
- `counter_throw()`: Attempt opportunistic throw during opponent strike. Requires `grip_level >= 1`.
R3. Exchange Resolution per Tick:
- **Grip Contest:** If an agent attempts a grip and succeeds contest check (`P(grip) = sigmoid(technique_diff + stamina_diff/15)`), `grip_level += 1`.
- **Throw Contest:** If `attempt_throw` is selected:
- Full Back Fall Probability: `P(ghand) = 0.12 * grip_level + 0.008 * (stamina_A - stamina_B)`. [gap-fill]
- Side/Hip Fall Probability: `P(vaza) = 0.25 * grip_level + 0.005 * (stamina_A - stamina_B)`. [gap-fill]
- If `ghand` succeeds: Direct Victory (`BOUT_END_GHAND`).
- If `vaza` succeeds: Agent scores +1 point (`points += 1`), `grip_level` resets to 0.
R4. Passivity and Penalties:
- Choosing `defend_sprawl` for 3 consecutive ticks adds 1 `passivity` warning.
- 3 passivity warnings award 1 point to opponent.
R5. Victory and Overtime:
- Instant win on `ghand`.
- At tick 30, if points are unequal, wrestler with higher points wins.
- If points are tied, 18-tick Golden Score overtime is played; first score wins. If still tied, wrestler with fewer passivity warnings wins.
## Edge cases
- **Tie at the end condition:** Overtime golden score followed by passivity penalty evaluation breaks all ties.
- **A player with no legal move:** Defend and stance actions are always valid.
- **Repetition / cycles:** Bounded by strict tick countdown (30 regulation + 18 overtime ticks max).
- **Simultaneous throw commitments:** Evaluated simultaneously; if both succeed throw checks, priority goes to higher roll differential.
- **Player elimination mid-round:** 1v1 bout terminates immediately upon direct throw.
- **Resource exhaustion:** Stamina reaches 0 floor, causing severe vulnerability to opponent throws.
- **Termination guarantee:** Guaranteed by finite tick limit and overtime decision tree.
## State representation
- **Persistent state:**
- `tick: int` (0..48).
- `agent_stamina: Tuple[float, float]` (0.0..100.0).
- `agent_grip_level: Tuple[int, int]` (0..3).
- `agent_points: Tuple[int, int]`.
- `agent_passivity: Tuple[int, int]`.
- `bout_status: enum {ONGOING, WIN_P0_GHAND, WIN_P1_GHAND, WIN_P0_POINTS, WIN_P1_POINTS}`.
- **Information set:** Fully observable public state.
## Action space
- `establish_jacket_grip()`: Precondition: `stamina >= 8`.
- `establish_sash_grip()`: Precondition: `stamina >= 12`, `grip_level >= 1`.
- `attempt_throw(technique: enum)`: Precondition: `grip_level >= 1`, `technique ∈ {PICH, KOOLANDAZ, LENGEH}`.
- `counter_throw()`: Precondition: `grip_level >= 1`, `stamina >= 15`.
- `defend_sprawl()`: Always valid.
## Playability notes
- **Single-player-vs-AI:** Highly playable. Player balances grip escalation with stamina conservation, timing explosive throws when AI stamina dips.
- **Group games:** Run as single-elimination tournament bracket.
- **Physical-skill model:** Grip torque, hip leverage, and balance recovery are abstracted into stamina-conditioned probability curves.
- **What is lost:** Physical jacket fabric friction and tactile foot-sweeping resistance.
## Implementation hints
Lightweight state machine. Implement stamina depletion on offensive actions and ensure `defend_sprawl` allows stamina recovery while risking passivity penalties.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Khorasan Choobeh Traditional Wrestling (Iran). (n.d.). Games of Countries. https://gamesofcountries.com/games/ir-iran/koshti-choobeh
@misc{gamesofcountries-ir-iran-koshti-choobeh,
title = {Khorasan Choobeh Traditional Wrestling (Iran)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/ir-iran/koshti-choobeh}
}