The Island
Take two rabbits. Drop them on a tiny island. Give them carrots, sunlight, and a bottomless appetite to reproduce. Call it a cute simulation. Call it a biology experiment. Call it a game. But under the fur and the twitching noses, Rabbit Island is a machine for asking one of the oldest questions in science: why do we ever cooperate when selfishness pays so well?
Every rabbit is an agent with simple rules. It hops. It eats. It gets hungry. It grows older. When it has enough energy and a willing partner, it makes a new rabbit. It carries a single gene for selfishness, a number from zero to one. The gene is inherited, mutated slightly, and tested by the island. The whole population is a living trial of the evolutionary game theory that underpins everything from kin selection to the tragedy of the commons.
The Theory
Evolution is not about the good of the species. It is not about the island. It is not even about the rabbit. It is about the gene. A gene that helps itself get copied into the next generation will spread, regardless of what it does to the body that carries it. Richard Dawkins called this the selfish gene, and the name is both perfect and misleading: the gene is not malevolent, but it is relentlessly self-promoting.
This creates a puzzle. If genes are selfish, why do rabbits ever share? Why do alarm calls, grooming, and resource sharing evolve in nature? The answer is that a gene can help copies of itself that live in other bodies. Saving a sibling is genetically cheaper than dying to save a child, but the math still works out because siblings share many genes. Helping a close relative is a form of helping yourself. That is kin selection, and it is the first force operating on Rabbit Island.
When the population is small and the rabbits are closely related, cooperation can emerge. A rabbit that eats a little less and lets a cousin finish a carrot might lose a tiny amount of energy, but if the cousin survives to reproduce, the shared genes survive too. The family wins. This is where the island first feels peaceful: a few warrens, shared meals, and population growing gently.
From Family to Clan
As the rabbits multiply, the island becomes crowded. Carrot patches that once seemed infinite are now contested. Resources are no longer the only constraint; neighbors are. The rabbits begin to sort themselves into clans or families of warrens that share territory and food. Related rabbits cluster because helping a relative is better than helping a stranger. The clans act like economic cooperatives: a rabbit may not eat every carrot it finds, but it may leave one for a sibling or a half-sibling, because the shared gene pool benefits.
This is where Rabbit Island becomes more than a screensaver. The clans are not scripted. They are an emergent pattern, the result of hundreds of small decisions made by individual agents that share genes. The family that balances reproduction and consumption survives longer. The family that overgrazes collapses. The island itself becomes the experiment, and the rabbits are the data.
The Game Theory
At its core, Rabbit Island is a repeated multiplayer game. Each rabbit has a strategy encoded in its selfishness gene. A very selfish rabbit tries to eat every carrot it sees. A less selfish rabbit may leave food for others. Each interaction is a tiny prisoner's dilemma: the selfish choice is the best individual choice in a single round, but if everyone is selfish, the group collapses. The repeated rounds change the math. In a repeated game, reputation and retaliation matter. A selfish rabbit that steals too much may find the local clan pushes back, or its own offspring inherit too much greed and starve when resources run low.
There is a second, more subtle game: the tragedy of the commons. Carrots are a shared resource. Every rabbit gains by eating one more carrot, but the cost of depletion is shared by the whole population. The dominant strategy is to eat now, but the equilibrium of everyone eating now is starvation. On the island, this plays out in real time. A selfish population can strip the island bare and crash. A cooperative population can maintain a steady harvest. The game is not fair; it is an iterated simulation of what happens when individual incentives collide with collective survival.
The Selfish Mutation
Then, one day, a mutation. A rabbit is born with a higher selfishness gene than its parents. It eats more. It shares less. It survives. It breeds. If its offspring inherit the trait, the gene pool begins to tilt. The selfish rabbits outcompete the cooperators in the short run because they take more than they give. They are, individually, the winners of the local game.
But the island has a long memory. The more selfish the population becomes, the fewer carrots remain. The warrens that once shared food now fight over scraps. The birth rate falls. The island enters a boom-and-bust cycle. The selfish gene, which won every local interaction, becomes the architect of the population's collapse. This is the paradox of the selfish gene: success at the individual level can be a failure at the system level. It is the same tension that appears in economics, ecology, and human societies. It is why the simulation matters.
Why This Matters
Rabbit Island is not a lecture. It is a feeling. You start with two rabbits and a sunset, and you end with a civilization that lives or dies by its own values. The importance of the project is that it makes abstract evolutionary forces concrete. Kin selection, repeated games, and the tragedy of the commons stop being textbook terms and become visible behaviors: a rabbit pauses at a carrot, a family cluster forms on the far shore, a selfish bloodline sweeps the island and then starves.
It also matters because the dynamics are genuinely unpredictable. The same starting conditions can produce a stable cooperative society, a boom-and-bust cycle, or a quick extinction. The gene mutation is random. The carrot regrowth is random. The encounters are random. The result is a complex system where large-scale order emerges from simple local rules, the hallmark of the most interesting phenomena in nature and society.
How to Play
Visit games.catgameresearch.net/rabbit-island.html to start the simulation. The controls are intentionally simple, but the behavior is anything but.
- Move the camera: W, A, S, D or the arrow keys. Hold to fly through the world.
- Look around: Left-click and drag, or right-click and drag depending on the camera mode.
- Zoom: Scroll wheel.
- Select a rabbit: Click on any rabbit. The HUD will show its energy, age, selfishness gene, and current state.
- Reset the simulation: Use the UI controls to start a new island with the same or different settings.
- Watch the population graph: The live plot shows population, births, and deaths over time. The shape of the curve tells you what kind of society the rabbits are building.
- Observe the sky: The time-of-day cycle is not cosmetic. Energy regen and feeding behavior are tied to the day-night rhythm.
Game Features
- GLTF-rendered rabbits and carrots with real-time instanced rendering and individual per-agent scaling.
- Genetic inheritance with mutation: each rabbit has a selfishness gene that evolves across generations.
- Kin- and clan-based behavior that emerges from local interactions, not hard-coded rules.
- Resource scarcity simulation with carrot regrowth, energy costs, and starvation.
- Live HUD and population graphs that update as the rabbits live, breed, and die.
- Day-night cycle affecting energy regen and activity.
- Interactive camera with keyboard and mouse controls.
Play It and Find Out
We built Rabbit Island because the best way to understand a game is to play it. The science is real, the math is real, and the rabbits are adorable. But the outcome is up to the gene pool.
Two rabbits. One island. Infinite carrots. What happens when a selfish rabbit comes along and taints the gene pool? Play the game and find out.