Formal Specification
Traditional Mongolian wrestling — Formal Specification
Mongolia · wrestling-combat · Confidence: low
← Back to game pageImplementation Suitability
Suitability: adaptation-requiredFidelity: probabilistic-proxyGap-fills: 2
# Mongol bökh — AI-optimized specification
## General idea
A two-agent combat grappling contest where opponents circle, engage in grip fighting (*zodog-shuudag* holds), and select tactical throwing techniques (trips, leg sweeps, hip throws, body lifts). Execution is resolved via a technique-matchup matrix and wrestler attributes (strength, balance, agility); any body contact other than the soles of boots touching the ground ends the bout instantly.
## Formal model
- **Abstraction taken:** The continuous physical wrestling match is discretized into structured grappling phases `t = 1, 2, ...`. Each phase consists of simultaneous grip selection and technique commitment, resolved through a stochastic fall checker.
- **Player/agent count:** 2 competitors per bout.
- **Turn model:** Simultaneous action selection per phase, resolved sequentially.
## Formal rules
R1. Setup: Both wrestlers start in the "Standing / Circling" state with `balance = 100` and `grip_status = NONE`.
R2. Grappling phases:
- Phase 1 — Circling & Hand-fighting: Agents choose an initial grip target: `{COLLAR, SLEEVE, BELT_CORD, SHUUDAG_LEG, NONE}`.
- Phase 2 — Grip established: Once grips are engaged, both agents simultaneously select one throwing technique from the traditional repertoire:
- `LOWER_LEG_TRIP` (*Kholiin makh*) — high speed, effective against heavy stances.
- `HIP_THROW` (*Gants khatuu*) — powerful counter to forward momentum.
- `INNER_LEG_SWEEP` (*Daa barih*) — leverages close body lock.
- `BODY_LIFT` (*Örgök*) — relies on pure strength advantage.
- `COUNTER_DEFENSE` (*Khavsarkh*) — sacrifices offense to restore balance.
R3. Stochastic resolution:
- Technique matchups interact to yield fall probabilities `P(A throws B)`, `P(B throws A)`, and `P(no fall, balance shift)`.
- [gap-fill] Matchup probabilities are weighted by relative agent attributes: `Strength`, `Agility`, and `Balance`.
R4. Fall detection:
- A fall occurs if either agent's balance drops to 0 or a high-amplitude throw succeeds.
- Ground contact by torso, knee, elbow, or head ends the bout immediately in the other agent's favor.
R5. Stalemates & Forced Grips (*Zolgokh*):
- [gap-fill] If no fall occurs after 10 grappling phases, the referee forces the wrestlers into a mutual right-side grip (*baruun zolgokh*), increasing fall probabilities by 3× to ensure rapid termination.
R6. Win/Loss/Draw: A single fall ends the bout (Win/Loss). Draws do not exist; the forced grip ensures a fall.
## Edge cases
- Tie at the end condition: Simultaneous double falls are resolved by declaring the wrestler whose body touched the ground first as defeated, or restarting the phase immediately.
- A player with no legal move: Not applicable — a defensive or offensive action is always available.
- Repetition / cycles: Bounded by R5 forced grip rule (*zolgokh*), which rapidly forces balance to 0.
- Simultaneous claims on the same resource: Simultaneous technique selections resolve through R3 probability matrix.
- Player elimination mid-round: Bout ends on first fall; in a 512-wrestler tournament, the loser is eliminated and the winner advances to the next *davaa*.
- Resource exhaustion: Not applicable.
- Termination guarantee: Guaranteed by the monotonically increasing fall probabilities under the forced grip phase (R5).
## State representation
- Persistent state: `balance_A` (0..100), `balance_B` (0..100), `grip_state` (struct of established grips), `phase_counter` (integer, 1..20), `forced_grip_active` (boolean).
- Per-phase ephemeral state: Secret technique choices `act_A` and `act_B`, sampled throw outcome.
- Observable information: Current grips, balance estimates, and wrestler attributes are public; current-phase technique choice is private until simultaneous reveal.
## Action space
- `select_grip(grip_type)`: Choose grip point on opponent's attire.
- `execute_technique(technique_enum)`: Choose throw or defense maneuver.
## Playability notes
- Single-player-vs-AI: Natural — 1 human wrestler vs. AI opponent in tactical rock-paper-scissors-plus-attributes combat.
- Group games (n>4): Tournaments scale cleanly as single-elimination brackets of 1v1 bouts (e.g. 512 wrestlers over 9 *davaa*).
- Physical-skill games: Parameterized via wrestler attributes `(Strength, Agility, Balance_Recovery)`.
- What is lost in the adaptation: The live physical resistance, leverage nuances, footwork on wet grass, and the ceremonial *devekh* eagle dance.
## Implementation hints
Simultaneous-move combat engine. Milestone 1: 5-technique rock-paper-scissors matrix with static balance tracking; Milestone 2: 9-round single-elimination tournament simulation generator.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Traditional Mongolian wrestling (Mongolia). (n.d.). Games of Countries. https://gamesofcountries.com/games/mn-mongolia/mongol-bokh
@misc{gamesofcountries-mn-mongolia-mongol-bokh,
title = {Traditional Mongolian wrestling (Mongolia)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/mn-mongolia/mongol-bokh}
}