Project made for Hack the Coast 2026

Last Call is a game-first crisis management game about running emergency shelter intake during a surge night (extreme cold/heat, displacement, etc.). The player makes irreversible intake decisions under scarcity, time pressure, and incomplete information. The point is the system: even “good faith” choices can produce unequal outcomes because constraints and delays don’t affect everyone equally.

• This is a game, not a tool, dashboard, or educational/awareness product.
• Scarcity is real (limited beds, staff capacity, supplies).
• Irreversibility: intake decisions cannot be undone.
• Incomplete information: player only sees plausible, partial traits.
• Delayed consequences: impacts show up later (often at night), not instantly.
• Equity/inclusion emerges mechanically, not via lectures or moral messaging.
• Failure should feel uncomfortable but fair (traceable to choices, not random punishment).

1. People arrive with partial, ambiguous information.
2. Player chooses one action: Admit / Deny / Delay.
3. Resources update immediately (beds/staff/supplies).
4. Consequences resolve later via a delayed event system.
5. The system trends toward collapse; player fights to keep it functional.

• Morning (Intake): fast admit/deny/delay decisions under pressure.
• Day (Strain): make a small number of system-level policy choices (no chores/micromanagement).
  • Examples: staff focus (intake vs care), rationing (strict vs generous), bed policy (overcrowd vs enforce).
  • Goal: choose where the system fails, not “fix everything.”
• Night (Resolve): no input; delayed outcomes trigger and system state updates.

• Third-person player character represents an intake coordinator.
• Movement is used to enforce time/bottlenecks (walking between stations), not to “solve” problems directly.
• Player does not perform chores (clean/cook/etc.) or hero actions.

• People are generated with visible traits (shown) and hidden traits (used for outcomes).
  • Visible examples: age band, presenting condition, group size, wait time.
  • Hidden examples: risk thresholds, vulnerability to delay, future system cost.
• Randomness is asymmetric: not all cases are “fair” or obvious.

• Outcomes are resolved later to prevent instant optimization and to make harm legible over time.
• Use a hybrid:
  • Decision-seeded events (probabilities scale with admitted profiles + policies + shelter stress).
  • Global random events (external shocks like delivery delays), with severity scaled by shelter stress.
• Outcomes are operational, non-graphic, and shelter-relevant (e.g., medical escalation, sleep disruption, illness spread, staff burnout, admin failures).

• Staff/volunteers exist as capacity modifiers (numbers + fatigue), not controllable units.
• No individual task assignment; player only sets high-level focus/policy.
• Overload can cause staff loss later; occasional arrivals are unpredictable.

• There is no “perfect win.” Success is fragile: keep the system functioning as long as possible.
• The game ends on system collapse (e.g., staff depleted, intake throughput < arrivals, queue overflow, critical capacity failure).
• No moral scoring; outcomes should feel earned and traceable.

• Tutorials or preachy dialogue about inequality/accessibility.
• Rewind/undo, “correct answer” moral choices, or comfort features.
• Complex economies (currency/income) or heavy micromanagement.
• Multiple modes/scenarios; keep scope intentionally small.