Formal Specification
Blue Wolf / Goat Carcass Grab — Formal Specification
Kazakhstan · animal-contest · Confidence: low
← Back to game pageImplementation Suitability
Suitability: adaptation-requiredFidelity: discretizedGap-fills: 5
# Kokpar — AI-optimized specification
## General idea
Kokpar is a high-intensity traditional equestrian team sport in which two four-rider mounted squads battle for possession of a 32-kilogram carcass (or certified synthetic mock carcass). Riders lean from their saddles to lift the carcass from the earth, gallop down the pitch while fending off opponents in high-speed scrums, and attempt to deposit the carcass cleanly into the opponent's goal ring (*sheñber*) or cauldron (*taiqazan*).
## Formal model
- **Abstraction taken:** The 200m x 70m dirt pitch is discretized into a 1D longitudinal grid of 7 zones: , where Zone 0 is the center spot, Zone is Team 1's goal ring (*sheñber*), and Zone is Team 2's goal ring. [gap-fill] Match time is discretized into discrete 5-second action ticks (total 240 ticks per 20-minute half); physical scrums and carcass possession contests are resolved via multi-agent strength/horsemanship attribute comparisons.
- **Player/agent count:** 2 teams (), each fielding exactly 4 mounted agents simultaneously.
- **Turn model:** Simultaneous tick-based discrete action selection across all 8 active riders.
## Formal rules
R1. **Pitch Geometry & Team State Initialization:**
- Zone layout: .
- Team 1 attacks toward Zone ; Team 2 attacks toward Zone .
- Each agent has attributes: Physical Strength , Equestrian Speed , and Wrestling Grip .
- Carcass state: `(zone, holder_id)` where `holder_id` is nullable.
- Match duration: Two halves of 240 ticks (total 480 ticks = 40 minutes active time).
R2. **Kickoff and Ground Retrieval:**
- At match start (and after every scored goal), the carcass starts unheld on the ground at Zone 0.
- Active riders sprint toward Zone 0; the first rider to enter Zone 0 initiates a pickup attempt `lift_carcass()`.
- [gap-fill] Pickup success probability: .
- If successful: Rider gains possession (`holder_id` ).
- If failed: The carcass remains loose on the ground in Zone 0.
R3. **Possession, Movement, and Passing:**
- On each tick, the carcass holder selects one action:
1. `advance(direction)`: Attempt to gallop into adjacent zone.
2. `pass_carcass(teammate_id)`: Pass to a teammate in the same or adjacent zone. Pass success probability: .
3. `hold_and_shield()`: Brace horse to absorb defensive contact.
- Non-holding riders select: `move_to_zone(target_zone)` or `engage_scrum(opponent_id)`.
R4. **Scrum Grapple Contest (*Doda*):**
- If one or more opposing defenders occupy the same zone as the carcass holder and select `engage_scrum`:
- [gap-fill] Scrum retention vs strip differential:
- If : A strip occurs; possession transfers to the strongest opposing defender in that zone.
- If a defender commits a high-risk illegal foul (whip strike, perpendicular horse ramming): the defender is assessed a 2-minute penalty (24 ticks in penalty box).
R5. **Scoring (*Salym Salu*) and Victory:**
- If an active holder enters the opponent's goal zone (Zone for Team 1, Zone for Team 2) and executes `score_goal()` without being stripped:
- The team scores +1 goal (*salym*).
- Play pauses, and the carcass is reset to Zone 0 for kickoff.
- At full time (480 ticks), the team with the higher score wins.
- [gap-fill] If tied at full time, a 10-minute (120-tick) sudden-death golden goal extra time is played. If still tied, a best-of-3 individual 1-on-1 horseback sprint shootout (*jekpe-jek*) decides the match.
## Edge cases
- **Tie at the end condition:** Explicitly resolved by R5 [gap-fill] (golden goal extra time, then 1v1 shootout).
- **A player with no legal move:** Not applicable; zone movements, passing, and scrum engagements are always available.
- **Repetition / cycles (and the loop-breaking rule):** Continuous scrums in midfield are bounded by the strict 480-tick game clock.
- **Simultaneous claims on the same resource or position:** In loose carcass scrambles, simultaneous pickup attempts are resolved by comparing agent composite attributes .
- **Player elimination mid-round; turn order repair:** If a rider is yellow-carded (2-minute box) or red-carded, their team plays shorthanded (3 vs 4) for the penalty duration; zone tick processing remains uniform.
- **Resource exhaustion (pieces, tokens, space, time):** 1 carcass in play; zone grid is fixed.
- **Termination guarantee: why the game provably ends:** A match is strictly bounded by 480 ticks plus at most 120 overtime ticks and a finite 3-round shootout.
## State representation
- **Persistent state:**
- `score`: `[team1_goals, team2_goals]`.
- `current_tick`: Integer .
- `team_rosters`: Array of player structs with persistent attributes.
- `penalty_timers`: Map of `player_id -> int`.
- **Per-tick ephemeral state:**
- `rider_zones`: Map of `player_id -> Zone (-3..+3)`.
- `carcass_state`: `{zone: -3..+3, holder: Nullable[PlayerID]}`.
- `active_actions`: Map of `player_id -> Action`.
- **Observable information:** Fully observable. Positions of all riders, carcass location/possession, scores, and game clock are public knowledge.
## Action space
- `move_zone(target_zone)`: Move to adjacent zone ().
- `lift_carcass()`: Pick up loose carcass in current zone.
- `pass_carcass(teammate_id)`: Pass to teammate within 1 zone.
- `engage_scrum(target_id)`: Defensive tackle/strip action.
- `score_goal()`: Deposit carcass into opponent goal ring.
## Playability notes
- **Single-player-vs-AI:** Natural. 1 human commanding a team squad (or riding as captain) against AI agents using zone positioning and coordinated blocking.
- **Group games (n>4):** Reduces cleanly to 2 commander agents issuing tactical zone and action directives to their 4-rider teams.
- **Physical-skill games:** Modeled via player athletic attributes .
- **What is lost in the adaptation:** Physical exhaustion of hauling a wet 33 kg carcass one-handed at full gallop, horse-to-horse torso collisions, and dust swirling in the arena.
## Implementation hints
- Implement as a discrete zone state machine with 7 nodes.
- AI team coordination can be implemented with role assignment: 1 Ball Carrier, 1 Lead Blocker, 1 Trailing Receiver, 1 Defensive Sweeper.
- First milestone: Implement 7-zone pitch, center kickoff sprint, single-rider zone advancement, goal scoring, and goal reset over 480 ticks.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Blue Wolf / Goat Carcass Grab (Kazakhstan). (n.d.). Games of Countries. https://gamesofcountries.com/games/kz-kazakhstan/kokpar
@misc{gamesofcountries-kz-kazakhstan-kokpar,
title = {Blue Wolf / Goat Carcass Grab (Kazakhstan)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/kz-kazakhstan/kokpar}
}