ANT-POCALYPSE
Designing a Game You Aren't Allowed to Play
The jam theme was Hands-Off: the player may only interact during a setup phase. To prevent a screensaver, setup carries the entire decision space under a strict 100-cell excavation budget—where every dig is a high-stakes trade.
Telemetry & Nest Profile
Sculpt an underground ant colony before the storm arrives. Once simulation starts, you can only watch.
Nest Operations Parameters
- Theme Constraint: Pure "Hands-Off" gameplay. All player actions occur strictly during the pre-run setup phase.
- Excavation Budget: Exactly 100 cells of dirt available to dig. Zero post-setup digging allowed.
- Pathfinding Architecture: Lightweight node swarms following dynamic
AStarGrid2Dtile paths. - Environmental Threats: Rainwater flooding upper galleries and escalating queen starvation timers.
- Diagnostic Telemetry: Precise timestamped postmortem event logs replacing generic lose screens.
- Development Time: Built from scratch in 48 hours for a game jam using Godot 4 & GDScript.
Design Statement
Ant-Pocalypse solves the primary failure mode of "Hands-Off" game design: avoiding a passive screensaver. Without tight constraints, the player's optimal strategy is always "excavate everything," eliminating decision-making. By restricting excavation to exactly 100 cells, every tunnel dug becomes a critical tactical trade-off.
The Setup Constraint
Without post-setup interaction, every choice must carry weight. A hundred cells of dirt turn excavation into pure tactical trade-offs.
Second Entrance vs. Depth
A second entrance provides a vital escape route and faster surface foraging, but every cell spent on a secondary shaft reduces the maximum depth your queen can retreat to.
Wide Gallery vs. Flood Speed
Wide two-tile galleries prevent ant bottlenecks and allow worker swarms to transport food in parallel. However, wide tunnels hold double the volume of rainwater and flood twice as fast.
Shallow Queen vs. Starvation
Placing the Queen chamber near the surface shortens food delivery trips, keeping her well-fed early on. But when the cloudburst strikes, a shallow chamber drowns long before deep burrows.
The Postmortem Engine
What turns a hands-off simulation from a one-time toy into an addictive tactical puzzle.
Generic Lose Screen
> YOU LOST.
> PRESS RESTART TO TRY AGAIN.
Played once. Provides zero feedback on why the nest failed or what layout adjustments to make.
Diagnostic Postmortem Log
> 1:15 Foraging route severed by water level
> 2:03 Queen starved in lower chamber
> DIAGNOSTIC COMPLETE: RE-ESTABLISH DRAINAGE
Played four times. Gives exact timestamped failure mechanics so players immediately know how to optimize setup.
Thin Simulation Architecture
Deliberately lightweight systems designed so the player reads a living swarm, not individual physics boxes.
Zero Physics Overhead
No heavy RigidBodies or physics collision solvers. Ants are plain Node2D entities walking pre-computed AStarGrid2D grid paths. Ants overlap freely, allowing hundreds of entities to run smoothly without frame drops.
16-Pixel Swarm Readability
At a 16-pixel tile resolution, the human eye reads collective swarm movement rather than individual unit jitter. Removing entity collision allowed us to polish water propagation and path readability instead.
Jam Pure Stack
Built entirely in 48 hours using Godot 4 and native GDScript. No external physics engines, third-party plugins, or heavy asset packages.