Towards Ecological Game Design

Josh Antolovic is a designer, artist, and researcher whose work centers on video games and natural systems.

Towards Ecological Game Design

2024-26

A framework, methodology, and public design guide for building responsive, ecologically alive natural environments in games, developed through a year of Unity prototyping, published, and pilot-tested with eighteen design students.

→ Version 1 of the Ecological Game Design Guide
→ Games for Good: Designing Virtual Ecosystems Across Scales
→ GDEX: Rethinking video game natural environments through algorithmic descriptions

Role: Solo (research, design, development, writing, and workshop facilitation)

Tools: Unity, system mapping & paper prototyping, custom guided-journaling app, Google Docs, PowerPoint, workshop facilitation & pedagogy

Approach: Research through and for design: the framework and methodology came out of what a year of prototyping taught me. Written for any designer working with natural environments in games, from a solo hobbyist to a system added mid-production on an already-shipping game.

The Question

Walk through an open-world game and you may witness something odd: what looks like lush landscapes that are, paradoxically, ecologically quiet. Soil holds up under endless farming. Rain is an animation laid over the camera. In a game like Minecraft, the world becomes but a field of resource cubes. Games are good at making nature lush and resourceful. They are rarely designed to make it behave like an ecosystem, with cause and effect that persists.

Wanderer above the Sea of Fog (Friedrich, 1818) with examples of grand landscapes within various AAA video games

That gap is where this project starts: what would it take for a game's natural environment to carry as much systemic weight as its player character does?

A handful of existing games gesture at this already. Far Cry 5's biome-painting tools. A Minecraft mod called Pollution of the Realms that lets fuel combustion poison plants and animals. However, it is rare for games to treat the environment as a first-class dynamic system the way a player character is treated. That's the space this guide works in.

Slide from Etienne Carrier’s 2018 GDC presentation on sub-biome recipes

Diagram from Pollution of the Realms’ webpage

Research Through and For Design

I came to programming through architecture and drawing rather than computer science. I took a four-month course on programming for artists and designers, and beyond that, learned by doing, more recently with LLMs doing a lot of the code-writing while I organized the ecosystem's overall structure. That division of labor is deliberate, and it mirrors what the guide teaches: the designer's job is system organization, with less emphasis on translating the system for computation. The project ran on Research through Design: prototypes came first, and the framework and methodology were built by paying close attention to what the prototypes taught.

Diagram of my RtD cyclical process

To keep from losing what I learned mid-build, I built a small guided-journaling app with three tracks: Studio (post-session reflection), Playtest (informal feedback from studio mates), and Field Trip (notes from time in real gardens and on trails, to compare against the virtual version). Everything compiles into one exportable document, which became the reference I kept returning to while writing the guide.

Diagram of the journaling effect on the RtD process

Screenshots of the guided journaling app and process

The Framework

The framework is the "what": a way of orienting toward a system before touching a game engine. It starts with scenario-based thinking, picking either a speculative game idea or a real natural system with visible cause and effect, then narrows to at least two spatial and temporal scales, different enough to contrast but close enough to interact.

Scenario mood board example

For my own prototypes I picked "gardenly" and "geologic." Gardenly sits close to human experience: seasonal cycles, nutrient flow, succession. Geologic is a harder zoom-out: erosion, sediment, continental drift, the sense that even permanent-looking features are mid-transformation. A single rainstorm operates at both scales at once, watering plants at the gardenly and carrying sediment at the geologic.

Diagram of spatial/temporal scales and their stretch of influence

Annotated gardenly scale

Annotated geologic scale

Example of spatial and temporal nested scale diagram

The framework also carries a smaller ladder underneath it: affordance (the environment lets the player act on it), responsiveness (the environment remembers and adapts), and agency (the environment can express itself back). Most games stop at affordance. The guide pushes toward responsiveness first.

Diagram of transactions between environment and gameplay

Affordance example

Responsiveness example

Agency example

The Methodology

The methodology is the "how," and it's deliberately analogue-first. A designer maps a system's inherent properties, then what affects each property, then trims for what actually improves the play experience, then locates points where a player or another system can intervene, written as simple pseudocode: if this, then that. All of it happens on paper before touching software.

Once systems start to overlap, I leaned on two planning tools to keep from losing track: version maps, which show how systems accumulate release to release, and feature tables, which lay systems out in rows against versions in columns. Neither shows everything on its own, so I used both.

Example of feature table

The Prototype: Sandbox Gardening Sim

The prototype came out of a specific irritation: reading about the slow progress of carbon-capture technology the same week mature trees on my street were cut down for new signage. I wanted a soil bed with plants growing and a data overlay, something that made environmental cause and effect visible the way a game UI makes anything visible.

v0.1 Sandbox Gardening Sim prototype, plant growth per ground tile’s soil moisture

I used Claude early on to sketch ecological system relationships as mermaid diagrams before writing any code, then to help structure the actual data and update logic in Unity as plants absorbed water from tiles between 0.3 and 0.7 moisture. The prototype grew weekly: color added to the tiles, then root-spread and moisture overlays, then a feature table roadmap for nine planned versions.

Assortment of ecological system maps made with Claude

I didn't finish all nine planned versions. What I did finish was a handful of distinct builds, pause/play/fast-forward controls that weren't on the original roadmap but felt right in the moment, and a finding: I'd lost track of the actual equations underneath the system. The fix became part of the method. I used Claude to generate a documentation site that separated designer-facing system logic from the underlying computation, and built a custom editor window to isolate and tune individual mini systems outside the live scene.

Various stills of v0.3 Sandbox Gardening Sim iterations

Sandbox Gardening Sim documentation made with Claude

Custom module editor with tunable parameters

Singled out module scene showcasing plant life cycle and spread

From Sketch to Guide

The guide didn't arrive fully formed. It started as a speculative methodology I called Non-human Algorithmic Personas & System Mapping, borrowing "persona" from UX design before I dropped the term entirely. I wanted the designer's attention on the system, not on an individual character standing in for it.

Initial methodology, “Non-human Algorithmic Personas & System Mapping”

The framework came into focus a semester later, presented at the Games for Good 2025 conference at Clark University and later published as "Designing Virtual Ecosystems Across Scales" in the Journal of Games, Self, and Society. Somewhere in assembling the actual guide, I realized there wasn't a separate "academic" and "accessible" version to write; the guide simply is the framework and methodology, formatted for use.

Nested spatial and temporal scale diagram

Version 1 lives on Google Docs now, after a short life as a worksheet and then a plain HTML page. It has an Instructions page, an Index, and five Parts: scenario building, an annotated scale diagram, system mapping, a physical prototype and rehearsal, and a show-and-tell close.

Ecological Game Design Guide Index

Ecological Game Design Guide Part 0

Ecological Game Design Guide Part 1

Ecological Game Design Guide Part 1, continued

Ecological Game Design Guide Part 2

Ecological Game Design Guide Part 3

Ecological Game Design Guide Part 4

Ecological Game Design Guide Part 5

The Pilot Workshop

In Spring 2026, I ran the guide as a two-and-a-half-week assignment in a collaborative studio course at OSU, eighteen upper-level students from visual communication, graphic, and industrial design. Three in-person sessions, two asynchronous while I was traveling for a conference.

Formation of pilot workshop groups

Groups picked a shared scenario, then split into three settings each: natural land, tended land, and a body of water. One group chose a zombie-survival game and split it into a bog, a farm, and a caldera they nicknamed a "volcano lake."

Setting inspiration slides from Part 0

The bog group had zombies infecting plants that a player treats with foraged herbs, a small closed loop of harm and recovery.

Student work example 1; Part 4 paper prototype, Scene 1

Student work example 1; Part 4 paper prototype, Scene 2

Student work example 1; Part 4 paper prototype, Scene 3

Student work example 1; Part 4 paper prototype, Scene 4

Student work example 1; Part 4 paper prototype, Scene 5

Student work example 1; Part 1 mood board

Student work example 1; Part 2 annotated nested scale diagram

The volcano-lake group split their system between a daytime farming and harvest cycle and a night cycle where zombies get blown off vents that also release minerals into the air.

Student work example 2; Part 2 annotated nested scale diagram

Student work example 2; Part 3 system map 1

Student work example 2; Part 3 system map 2

Running it once surfaced various insights: students skimmed the guide rather than reading it closely, reached for digital tools over paper despite my encouragement, and needed real-time help narrowing scope more than I'd planned for. All of it, along with what worked, feeds directly into Version 2.

Prototype Showcase

The same spring as the workshop, a revised version of the prototype had its first fully public showing: at the ACCAD Open House in April 2026, I demoed the sandbox gardening simulator alongside the guide and the workshop outcomes.

Sandbox Gardening Simulator Prototype

What’s Next

The clearest gap in Version 1 of the guide is that it only teaches one path: starting a system from a blank scenario. The guide doesn't yet have a second path for a designer adding a system into a game that's already in production, working around existing constraints instead of an open field. That's the next concrete addition.

Beyond that, a short list of directions I want to keep pulling on: evaluative tools for reading ecological depth in existing games, prototyping toolkits that push designers toward ecological systems the way platformer toolkits push them toward jumping and gravity, further testing of the guide with real data collection, and a careful, separate look at anthropomorphic agency, giving the environment a literal voice.

@joshantolovic

Index Next →