Formal Specification
Handkerchief drop circle chase — Formal Specification
Azerbaijan · chase-tag · Confidence: medium
← Back to game pageImplementation Suitability
Suitability: adaptation-requiredFidelity: probabilistic-proxyGap-fills: 2
# Dəsmalaldı — AI-optimized specification
## General idea
A "drop the handkerchief"-family circle chase whose distinctive mechanic is the asymmetric race it triggers: once the targeted seated player notices the dropped handkerchief, both the Runner and the newly-risen pursuer are racing to the *same physical goal* — the vacated seat — from different points on the circle, so the outcome depends on relative position and speed at the moment of detection, not a direct tag contest. A captured Runner also serves a one-round "prison" penalty rather than simply losing a point.
## Formal model
- **Abstraction taken:** The stealth of the drop and the tactile/peripheral detection of the handkerchief are abstracted into a stochastic per-tick detection check for the targeted player only, since the entry describes both qualitatively without numeric models.
- **Player/agent count:** 1 Runner, M Seated players, M in [7, 19].
- **Turn model:** Synchronous ticks once the chase begins; the drop itself is a discrete triggering event chosen by the Runner during their circuit.
## Formal rules
R1. Setup: M seated players form a circle facing inward, hands in laps, forbidden from looking backward (entry §1.1-1.2, `Geri baxmaq olmaz!`); one Runner is chosen by counting rhyme and starts outside the circle holding the handkerchief (entry §1.3).
R2. Circuit and drop: the Runner moves clockwise around the circle's perimeter each tick and, at a moment of their choosing, submits `drop_handkerchief(target_id)`, placing it directly behind one seated player without a detectable sound or visual cue (entry §2.1-2.3). This sets `holds_handkerchief[target_id] = true` (hidden to all seated players until detected) and the Runner continues moving clockwise past that point.
R3. Detection: each tick after the drop, the targeted player alone resolves a stochastic `detection_check()` (entry §3, "feel behind their backs") [gap-fill: the entry does not quantify detection speed or the effect of the Runner's stealth on it; a flat per-tick detection probability is the minimal faithful reading, since no numeric relationship is given]. Other seated players have no handkerchief to detect and take no action.
R4. Chase trigger: the instant the target's `detection_check()` succeeds, they immediately stand, snatch the handkerchief, and begin `sprint(clockwise)` in pursuit from their seat position (entry §3.2); simultaneously the Runner continues `sprint(clockwise)` toward the now-vacated seat (entry §3.3). Both agents' positions at the moment of detection determine the race geometry.
R5. Resolution — Runner wins the race (entry §4.1): if the Runner reaches and sits in the vacated seat before being tagged, they are safe and become a seated player at that spot; the pursuer (still holding the handkerchief) becomes the new Runner for the next round.
R6. Resolution — pursuer wins the race (entry §4.2): if the pursuer tags the Runner with the handkerchief before the Runner reaches the seat, the Runner is captured (`tutuldu`) and must sit in the center ("prison," `hambal yeri`) for exactly one subsequent round, after which they rejoin the seated circle; the pursuer retains the Runner role for the next round.
R7. Match end: the entry specifies no win condition, describing the activity as "played in continuous rounds" (entry §5) [gap-fill: treated honestly — there is no formal terminal state, matching the entry's own framing rather than inventing a score threshold].
## Edge cases
- Tie at the end: not applicable — there is no scoring system or terminal state to tie at (R7).
- A player with no legal move: not applicable — the Runner always has `move`/`drop_handkerchief` available; the targeted seated player always has the passive `detection_check()` available; once racing, both racers always have `sprint` available.
- Repetition / cycles: possible and not structurally prevented — a Runner could in principle circle indefinitely without dropping, or the detection probability could keep a chase from ever triggering for many circuits; this is consistent with the entry's own framing of continuous, non-terminating play (R7) rather than a flaw.
- Simultaneous claims on one position: not applicable — exactly one seated player is targeted per drop (R2); the vacated seat is a single contested destination but R5-R6 resolve to exactly one outcome per race.
- Elimination mid-round and turn-order repair: a captured Runner (R6) is not eliminated from the game, only benched in the center for one round; no other player's state is affected, and they rejoin the seated pool afterward.
- Resource exhaustion: not applicable — no consumable resource; the handkerchief and circle are reused every round.
- Termination guarantee: **not guaranteed.** The entry gives this game no win condition and describes it explicitly as continuous play (§5), matching the pattern of comparable "no formal endpoint" games in this batch. `termination_guaranteed: false` reflects this honestly rather than inventing a stopping rule the entry does not support.
## State representation
- Persistent (session) state: `runner_id`, `round_number`, `prison_occupant` (if any, with a `rounds_remaining` counter, R6).
- Per-round hidden state: `holds_handkerchief[target_id]` — known only to the Runner and, once detected, the target themselves.
- Per-chase ephemeral state: `runner_position`, `pursuer_position`, `vacated_seat_position`.
- Observable information: all seated players other than the target have no information about who is targeted (hidden state); once a chase begins (R4), positions become fully observable to everyone.
## Action space
- `move(clockwise)`: the Runner only, circuit phase; precondition: no chase currently active.
- `drop_handkerchief(target_id)`: the Runner only; precondition: circuit phase, `target_id` is a seated player.
- `detection_check()`: the targeted player only (automatic/passive); precondition: `holds_handkerchief[self] = true`, chase not yet triggered.
- `sprint(clockwise)`: the Runner and the pursuer only, once a chase is triggered (R4).
## Playability notes
- Single-player-vs-AI: natural on either side — human seated player detecting an AI Runner's stealth drops, or human Runner managing drop timing and stealth against an AI-controlled seated circle.
- Reducing M > 4 to 2 agents: mechanically degenerate — with very few seated players the drop's stealth element becomes trivial (near-certain identification of who was targeted), removing the core detection challenge.
- Physical-skill games: real tactile detection, sprint-from-seated-start reaction, and stealth movement are the entire game and are collapsed into the detection stochastic check (R3); no genuine peripheral sensitivity or agility is modeled.
- What is lost: the actual tactile surprise of discovering the handkerchief, the traditional clapping and singing that accompanies the circuit, the comic tension of the teasing "fake drop" variant, and the social ritual of the "prison" seat.
## Implementation hints
State is small (a hidden target flag, positions during a chase, a prison timer); the real design difficulty is calibrating the detection-probability curve and whether/how Runner stealth modifies it, neither of which the source quantifies. First milestone: implement the drop-detect-chase loop (R2-R6) with a flat detection probability to validate the race-to-the-seat resolution and the prison penalty, then add stealth-sensitivity.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Handkerchief drop circle chase (Azerbaijan). (n.d.). Games of Countries. https://gamesofcountries.com/games/az-azerbaijan/desmalaldi
@misc{gamesofcountries-az-azerbaijan-desmalaldi,
title = {Handkerchief drop circle chase (Azerbaijan)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/az-azerbaijan/desmalaldi}
}