Skip to main content
Formal Specification

Leapfrog — Formal Specification

Bulgaria · strength-endurance · Confidence: medium

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: abstractedGap-fills: 3
# Preskochikkobila — AI-optimized specification ## General idea A sequential queue-based athletic leapfrog game with progressive hurdle heights. Agents execute vaults over stationary base agents; successful jumps advance hurdle difficulty and extend the sequential hurdle chain, while clearance failures incur immediate role demotion to the base position. ## Formal model - **Abstraction taken:** Gymnastic vault physics (run-up velocity, hand placement, hip elevation, leg spread) is modeled as a stochastic clearance check evaluated against hurdle height tier H∈{1,2,3,4} and agent athletic attributes. - **Player/agent count:** N agents (N≥4). - **Turn model:** Strictly sequential queue traversal. ## Formal rules R1. Setup: - Hurdle height tier H=1 (H∈[1,4]). - One agent is designated as Base Hurdle B. - Remaining N−1 agents form queue Q=[A1,…,AN−1]. R2. Vault Action: - Active agent at head of queue Aactive=pop_front(Q) executes `attempt_vault(target_hurdle)`. - [gap-fill] Clearance probability: P(Clear)=max⁡(0.15,AgilityA⋅0.8+JumpPowerA⋅0.3−0.20⋅H). R3. Outcome Resolution: - Success (Clear==True): - Agent Aactive lands cleanly and appends to the hurdle array (or queue rear). - If all hurdles in the active sequence are cleared: Aactive scores +H points. - Failure / Fault (Clear==False): - Vault fails. Aactive is immediately demoted to become the new Base Hurdle B. - Previous base hurdle Bold joins the rear of queue Q. - Hurdle height resets to H=1. R4. Level Escalation: - When all agents currently in queue successfully complete vaults at height H without any fault, height tier escalates: H←min⁡(4,H+1). R5. Level 4 Special Mechanic (Hat Balance / *S kapichka*): - At H=4, a stability check is added: P(HatStay)=PrecisionA⋅0.7. - If hat falls, vault counts as a fault per R3. R6. [gap-fill] Session Victory: - Game terminates after Kmax=10 full queue cycles. - The agent with the highest accumulated clearance points (or fewest rounds spent as base hurdle B) is declared the winner. R7. [gap-fill] Tiebreaker: If points are tied, the tied agents execute a sudden-death Level 4 vault; first to fail loses. ## Edge cases - Tie at the end condition: Fully resolved by R7 sudden-death Level 4 vault trial. - A player with no legal move: Not applicable — vault action is always legal on agent's turn. - Repetition / cycles: Bounded: Each cycle either advances height H or triggers a fault that resets H and rotates base B; Kmax queue cycles enforce absolute session termination. - Simultaneous claims on the same resource or position: Queue order is strictly sequential; exactly one vaulter jumps at any instant. - Player elimination mid-round; turn order repair: If an agent disconnects from queue Q, queue length |Q| decrements without disrupting sequential order. If base B disconnects, front of Q becomes the new base. - Resource exhaustion: Not applicable — physical hurdle entities are persistent and reusable. - Termination guarantee: Enforced by finite session limit Kmax and strictly positive failure probabilities at higher tiers (H≥3). ## State representation - Persistent state: Cumulative agent points [Score1,…,ScoreN], agent physical attributes (Agility,JumpPower,Precision), session cycle counter k∈[1,Kmax]. - Per-vault ephemeral state: Current height tier H∈[1,4], active base agent ID B, queue array Q, vault outcome boolean. - Observable information: Fully observable — all agent attributes, current height H, queue order, and scores are public knowledge. ## Action space - `attempt_vault()`: Precondition: Caller is at head of queue Q, base agent B is set. ## Playability notes - Single-player-vs-AI: Natural — Human controls one vaulter timing takeoff and jump commitment, competing against AI agents in the queue to maintain high difficulty tiers. - Group games (n>4): Natively designed for 4≤N≤12; scalable to any N by queue length extension. - Physical-skill games: Dynamic vaulting power is mapped to scalar jumping and precision attributes. - What is lost: The physical impact of hands pressing against shoulders and the mid-air spatial sensation of spreading legs over a human back are abstracted into probabilistic clearance checks. ## Implementation hints Extremely simple and lightweight state machine. Queue processing is O(1) per turn. Suggested first milestone: implement the N-agent circular queue loop with height escalation logic and verify that height tiers cycle realistically based on clearance probabilities.

Download as JSON

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

Download JSON
Cite this page

Leapfrog (Bulgaria). (n.d.). Games of Countries. https://gamesofcountries.com/games/bg-bulgaria/preskochikkobila

@misc{gamesofcountries-bg-bulgaria-preskochikkobila,
  title = {Leapfrog (Bulgaria)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/bg-bulgaria/preskochikkobila}
}