Research Backtest — MCP Tools

Drive the research backtester from an AI agent via the EventTrader MCP server. Add the server to your MCP client config, then call the tools below.

Tools

• et_research_backtest_themes — list selectable themes.
• et_research_backtest_models — list algo models (+ experimental caveats).
• et_research_backtest_run — start a run; returns run_id + the progress WS URL.
• et_research_backtest_runs — list recent runs.
• et_research_backtest_get — fetch one run's full JSON (summary + rows).
• et_research_backtest_leaderboard — ranked runs + launched cards.
• et_research_backtest_launch_rally_card — launch a basket as a Rally Card (auth; $0.99 fee).

Example agent prompt

"Backtest a DeFi + AI themed basket with a 24h +5% spike gate over a 2-day hold,
 then show me the top basket by vs-BTC return."

→ et_research_backtest_run(themes=["defi tokens","ai tokens"],
       spike_window_hours=24, spike_up_min_pct=5, hold_days=2)
→ et_research_backtest_get(run_id=...)

Note: the odiv/oboss models are experimental research (gross signal real, net-of-cost negative). The tools surface that caveat; never present them as tradeable.