Documentation Hub
Engineering the Survival
of Quantitative Agents.
A real-time, genetic-algorithm-based trading simulation. 14 autonomous agents compete on live Binance data, evolving through competition and survival of the fittest.
Arena Records
Interactive Strategy Preview
The Scalper
High-velocity agents designed to capture micro-trends. They utilize short-window indicators and high leverage to compound small gains.
The Hybrid
Balanced agents that mix momentum following with mean reversion. They wait for stronger confirmation than scalpers but trade more frequently than swing bots.
The Swing Trader
Patient, long-term agents that require multi-indicator alignment. They aim for major pivot points and large price swings.
Engineering Philosophy
Extreme Efficiency
Built entirely in Vanilla JS without heavy frameworks. This ensures minimal latency between WebSocket message arrival and trade execution. Indicators are calculated in O(1) time using incremental smoothing.
Evolutionary Alpha
The system does not search for a static "perfect" strategy. It assumes the market is constantly changing. By killing weak performers and breeding mutations of the strong, the Arena adapts to shifting volatility and trends.