Skip to main content
Formal Specification

Mud Arena Folk Wrestling — Formal Specification

Pakistan · wrestling-combat · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 2
# Pehlwani — AI-optimized specification ## General idea A two-agent grappling combat contest simulating traditional mud-arena folk wrestling. Agents transition through standing clinch, takedown attempts, ground scrambling, and bridging defense, with each phase modeled as a Markov decision process parameterized by wrestler stamina, technical skill, and strength. The bout terminates immediately upon pinning both of the opponent's shoulders flat to the ground. ## Formal model - **Abstraction taken:** Continuous physical grappling is abstracted into a discrete positional state machine: `{neutral_standing, dominant_clinch, grounded_top, grounded_bottom, bridging_defense, pinned}`. - **Player/agent count:** 2 agents (`Pehlwan_A`, `Pehlwan_B`). - **Turn model:** Simultaneous tactical commitments per discrete grappling round (round length = 5 seconds of simulated time; max 180 rounds = 15 minutes). ## Formal rules R1. Setup: Both agents start in `neutral_standing` with `stamina = 100`, `fouls = 0`. R2. Standing Phase: - Available actions: `{collar_tie, leg_dive_bagal, hip_toss_dhobi, sweep_taqla, disengage}`. - Simultaneous choices resolve via a technique-counter matrix weighted by `(strength, technique)`. - Successful takedown transitions one agent to `grounded_top` and the other to `grounded_bottom`. R3. Ground Control Phase: - Top agent actions: `{half_nelson, waist_lock, bridge_breaker, maintain_control}`. - Bottom agent actions: `{stand_up_escape, roll_reversal, bridge_defend}`. R4. Pin Fall Resolution (*Chit*): - If top agent executes `pin_press` while bottom agent is in `grounded_bottom` or breaks `bridging_defense`, a pin-fall check resolves: `P(pin) = (strength_top * stamina_top) / (defense_bottom * stamina_bottom + eps)`. - If pin succeeds: Bottom agent transitions to `pinned` state. Top agent wins the bout immediately. R5. Stamina Drain: Each high-power throw or defensive bridge consumes stamina (`stamina -= 5..15`). Low stamina heavily penalizes pin defense. R6. [gap-fill] Bout duration cap: Max 180 rounds. If no pin-fall occurs, the agent with higher accumulated control-time points is awarded victory. R7. [gap-fill] Fouls: Illegal concussive actions incur an immediate foul (`fouls += 1`). 2 fouls result in disqualification. ## Edge cases - **Tie at the end condition:** Handled deterministically by accumulated ground-control points at round 180 (R6). - **A player with no legal move:** Defensive posturing (`maintain_control` or `stand_up_escape`) is always available. - **Repetition / cycles:** Bounded by the 180-round clock and monotonic stamina depletion making late-round pin-falls statistically inevitable. - **Simultaneous claims on the same resource or position:** Bilateral techniques resolve synchronously via the transition matrix. - **Player elimination mid-round:** 1v1 bout terminates on first pin or foul-out. - **Resource exhaustion:** Stamina regenerates slowly during neutral resets but depletes during active struggles. - **Termination guarantee:** Guaranteed by finite round horizon (180 rounds) and absorbing `pinned` state. ## State representation - **Persistent state:** `bout_state in {neutral_standing, dominant_clinch, grounded_top, grounded_bottom, bridging_defense, pinned}`, `stamina_A`, `stamina_B`, `control_points_A`, `control_points_B`, `round_counter in {1..180}`. - **Per-round ephemeral state:** `action_A`, `action_B` (simultaneous hidden reveal), transition result. - **Observable information:** Current positional state, stamina bars, and score are fully observable. ## Action space - `select_grapple_action(action_id)`: Select one legal technique corresponding to the active positional state. ## Playability notes - **Single-player-vs-AI:** Natural; human player manages stamina expenditure while hunting for high-percentage takedown-to-pin combinations against an AI wrestler. - **Group games (n>4):** Tournament bracket format (e.g. 8-wrestler single-elimination dangal). - **Physical-skill games:** Leverage, balance, and core power are abstracted into the three-attribute vector `(strength, technique, endurance)`. - **What is lost:** The sensation of cool, turmeric-scented mud on warm skin, the friction of clay-coated hands slipping on biceps, and the roar of the drum at ringside. ## Implementation hints Low complexity. The 6-state Markov chain with 4–5 actions per state can be implemented in a few dozen lines of transition code. First tune the stamina degradation curves so that matches naturally build toward a decisive climax around rounds 50–100.

Download as JSON

Download the full game corpus entry for this game as structured JSON.

Download JSON
Cite this page

Mud Arena Folk Wrestling (Pakistan). (n.d.). Games of Countries. https://gamesofcountries.com/games/pk-pakistan/pehlwani

@misc{gamesofcountries-pk-pakistan-pehlwani,
  title = {Mud Arena Folk Wrestling (Pakistan)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/pk-pakistan/pehlwani}
}