Principle:HKUDS AI Trader Performance Visualization
| Knowledge Sources | |
|---|---|
| Domains | Data_Visualization, Performance_Analysis |
| Last Updated | 2026-02-09 14:00 GMT |
Overview
A data visualization technique that produces multi-metric comparison charts for evaluating trading agents against each other and market benchmarks.
Description
Performance Visualization creates standardized multi-subplot charts comparing the rolling performance metrics of multiple trading agents. Each chart contains four subplots showing the time evolution of Cumulative Return, Sortino Ratio, Volatility, and Maximum Drawdown. Agents are plotted as colored lines, with an optional market benchmark (e.g., QQQ for US stocks, SSE-50 for A-shares) shown as a dashed reference line.
This visualization enables at-a-glance comparison of agent performance over time, revealing periods of outperformance and underperformance relative to both other agents and the market.
Usage
Use this principle after computing portfolio values and rolling metrics for all agents in a comparison. The visualization is the final output of the multi-agent comparison workflow.
Theoretical Basis
# Pseudocode for multi-agent performance visualization
fig, axes = create_subplots(1, 4) # CR, SR, Vol, MDD
for metric, ax in zip(["CR", "SR", "Vol", "MDD"], axes):
for agent_name, df in agent_data.items():
ax.plot(df["date"], df[metric], label=agent_name)
if benchmark:
ax.plot(benchmark["date"], benchmark[metric], "--", label="Benchmark")
ax.legend()
save_as_pdf(fig, output_path)
Four key metrics visualized:
- CR (Cumulative Return): Shows wealth accumulation over time
- SR (Sortino Ratio): Shows risk-adjusted return focusing on downside
- Vol (Volatility): Shows expanding annualized volatility
- MDD (Maximum Drawdown): Shows worst peak-to-trough decline trajectory