Skip to main content
Formal Specification

Greasy Pole Climbing — Formal Specification

Kosovo · strength-endurance · Confidence: high

← Back to game page

Implementation Suitability

Suitability: adaptation-requiredFidelity: abstractedGap-fills: 4
# Ngjitja në shilor — AI-optimized specification ## General idea A turn-based sequential climbing race up a 10-meter vertical greased mast. On each tick of an ascent attempt, an agent chooses an exertion level (cautious shimmy vs aggressive surge), trading off height progression against exponential grip-fatigue and slippage risk. ## Formal model - **Mast Discretization:** The mast height Htotal=10.0 meters. - **Agent State:** Current height h∈[0,10.0], stamina E∈[0,100], accumulated grease friction coefficient μ∈[0.1,0.5]. - **Turn model:** Sequential attempts (2 official attempts per agent). Each attempt is simulated as a discrete sequence of 1-second climb ticks until summit is reached, stamina is exhausted, or a slip occurs. ## Formal rules R1. **Initial Setup:** N agents. Target height H=10.0 m. Each agent i has grip skill Gi∈[1,10] and stamina capacity Emax=100. Attempt limit per agent K=2. R2. **Tick Action Selection:** On tick t, the active agent selects a climbing pace action: - `CAUTIOUS_SHIMMY`: Height advance Δh=0.30 m, stamina cost ΔE=3.0, base slip probability pslip=0.05. - `STEADY_CLIMB`: Height advance Δh=0.60 m, stamina cost ΔE=7.0, base slip probability pslip=0.15. - `EXPLOSIVE_SURGE`: Height advance Δh=1.00 m, stamina cost ΔE=15.0, base slip probability pslip=0.35. R3. **Slip and Fall Resolution:** [gap-fill] - Net slip probability: Pslip=min⁡(0.95,pslip×(1.0+100−E50)×(1.0−Gi20)) - Sample ξ∼𝒰(0,1). - If ξ<Pslip (SLIP): - Agent slips down Δhdrop=min⁡(h,2.0 m) [gap-fill]. If h−Δhdrop≤0, the attempt ends with final height hfinal=0. - Otherwise (SUCCESS): - h←h+Δh, E←max⁡(0,E−ΔE). R4. **Attempt Termination:** An attempt terminates when: - h≥10.0 m (Summit reached! Attempt time recorded as Tattempt=t seconds). - E==0 (Exhaustion: agent slides to ground; height recorded as peak height reached). - t≥60 seconds (Time limit reached; recorded as current height). R5. **Tournament Outcome:** - Primary metric: Fastest summit time Tmin(i). - Secondary metric (if no summit reached): Maximum peak height Hmax(i). R6. **Tie-Breaker:** [gap-fill] If two agents have identical summit times (or peak heights), their second official attempt breaks the tie. ## Edge cases - **Tie at the end condition:** Handled explicitly by R6 (secondary attempt comparison; if still tied, a sudden-death single attempt is run). - **A player with no legal move:** Not applicable; all three pace actions are valid on every tick until attempt termination. - **Repetition / cycles:** Impossible; stamina decreases monotonically by at least 3.0 each tick, guaranteeing attempt termination in at most 34 ticks. - **Simultaneous claims on the same resource or position:** Not applicable; attempts are conducted sequentially in individual time trials. - **Player elimination mid-round / turn order repair:** If an agent times out on action submission, [gap-fill] an automatic `CAUTIOUS_SHIMMY` is executed. - **Resource exhaustion:** Stamina depletes to 0, which automatically concludes the active attempt. - **Termination guarantee:** Each attempt lasts at most 34 ticks (stamina depletion) or 60 seconds. With N agents and 2 attempts each, the tournament provably concludes in finite time. ## State representation - **Persistent state:** - `agent_records`: Array of N structs containing best summit time and peak height across attempts. - `active_agent_idx`: Integer ∈[0,N−1] - `current_attempt`: Integer ∈[1,2] - **Per-tick ephemeral state:** - `current_height`: Float ∈[0.0,10.0] - `current_stamina`: Float ∈[0.0,100.0] - `elapsed_time`: Integer (seconds) - **Observable information:** Fully observable to all agents. ## Action space - `climb_step(pace)`: parameter `pace` ∈{CAUTIOUS_SHIMMY,STEADY_CLIMB,EXPLOSIVE_SURGE}. ## Playability notes - **Single-player-vs-AI:** Natural. A human player attempts to optimize climb pace against AI leaderboards. - **Group games (n>4):** Reduces directly to an asynchronous leaderboard time trial. - **Physical-skill games:** Modeled as a stochastic resource-management problem balancing pace versus slip variance. - **What is lost in the adaptation:** The visceral friction heat, muscle burning, and crowd roar on the alpine meadow. ## Implementation hints - Pure Markov Decision Process (MDP) with deterministic cost/gain and stochastic slip transitions. - Can be solved analytically via dynamic programming to find the optimal pace policy π∗(h,E) for any given grip skill G. - First milestone: Implement single-attempt loop and plot height vs time trajectory for constant-pace strategies.

Download as JSON

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

Download JSON
Cite this page

Greasy Pole Climbing (Kosovo). (n.d.). Games of Countries. https://gamesofcountries.com/games/xk-kosovo/ngjitja-ne-shilor

@misc{gamesofcountries-xk-kosovo-ngjitja-ne-shilor,
  title = {Greasy Pole Climbing (Kosovo)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/xk-kosovo/ngjitja-ne-shilor}
}