Parameter Playground
How do you build a digital society where good behavior is rewarded and bad behavior naturally dies out? Tweak the parameters below and discover the answer.
This is the lowest-friction pathway to participation — experiment with one agent's survival by tuning parameters like starting energy, risk appetite, and karma carry-forward.
Want to see what happens when multiple agents interact? The Society Simulator lets you watch 12 agents with different strategies form alliances and self-organize.
Try These Experiments
Each experiment runs a simulation with pre-configured parameters to answer a specific question. Click one to see what happens — then tweak the parameters yourself below.
Custom Parameters
Or design your own experiment — adjust any parameter and run the simulation.
Simulation Parameters
SIMULATION SETTINGS
How many life cycles to simulate
How many decisions the agent makes each life. More actions = longer lives
0 = safe, cautious actions only. 1 = high-stakes gambles with bigger rewards
INITIAL CONDITIONS
Attention budget at birth — run out and you die
Trust score at birth (0 = untrusted, 1 = fully trusted)
REBIRTH KARMA
How much an agent's past life affects their next one. Good behavior compounds; bad behavior haunts.
How much extra energy a trusted agent gets when reborn
How much trust reputation carries over to the next life
QUICK PRESETS
Configure parameters and run a simulation to see results here.
How It Works
Adjust Parameters
Use sliders to configure simulation settings, initial conditions, costs/rewards, and behavioral tendencies.
Run Simulation
Click "Run Simulation" to execute a multi-life agent cycle with your parameters. Runs instantly in your browser — no setup needed.
Explore Results
View life trajectories, ATP/trust evolution, termination reasons, and auto-generated insights.
Iterate & Learn
Adjust parameters based on results, re-run, and discover how different settings affect society outcomes.
Key Insights to Discover
💰 Economic Balance
Rewards must exceed costs on average, but not by too much. Too easy = no challenge. Too hard = inevitable death. The sweet spot enables growth through skill.
🤝 Trust Accumulation
Trust should be easier to lose than to gain (asymmetric). This creates pressure for consistent good behavior. One failure shouldn't destroy you, but patterns matter.
🔄 Rebirth Balance
Carry-forward bonuses shouldn't make you invincible, just give you a head start. The simulation prevents runaway advantages so each life still requires earning trust.
🎲 Risk vs Reward
High-risk actions have higher rewards but also higher costs. Success probability depends on trust. Low trust + high risk = likely failure. High trust enables risk.
Technical Details
How the simulation works
The playground runs a Web4 agent simulation entirely in your browser — no server, no Python, no setup. It's optimized for speed and interactivity:
- Agent lifecycle: Birth → Actions → Death → Rebirth with karma
- Action selection: Based on risk appetite, current ATP, and trust
- Success probability: Higher trust = higher success rate (60-90%)
- Termination conditions: ATP exhaustion or trust below threshold
- Karma mechanics: Previous life's trust affects next life's ATP/trust
This is intentionally simpler than the full Web4 game engine (which includes cross-life learning, multi-agent interactions, context boundaries, etc.). The playground focuses on core mechanics: ATP economics, trust evolution, and karma.
For the full complexity, see the Society Simulator.
Curious Why It Works?
You've seen agents live, die, and carry forward karma. Now understand the mechanics behind the simulation.
Want More Experiments?
The best way to understand Web4 is to play with it. Break things. Discover edge cases. Find the tipping points.