Qc Backtest Master

QuantConnect automated backtest pipeline — submit local strategies, compile, execute, monitor with early-stop, and download results in one command.

Automate the full QuantConnect backtest lifecycle from local code to cloud results, without ever opening the web UI.

What this skill does

When the user asks to backtest a strategy, run a backtest, or submit code to QuantConnect, follow this workflow:

1. Submit & Compile

Upload the user's local .py strategy file to a QuantConnect cloud project, replacing any existing main.py. Trigger compilation and report any errors with line numbers.

python3 submit_backtest.py \
  --main-file <path_to_strategy.py> \
  --name "<backtest_name>" \
  --project-id $QC_PROJECT_ID

If the strategy has auxiliary data files (.json, .txt, .csv), upload those too.

2. Monitor with Early-Stop

Poll backtest progress in real time. If drawdown exceeds the user's threshold after 20% progress, automatically delete the backtest to save compute time.

python3 monitor_backtest.py <backtest_id> \
  --max-dd <threshold> \
  --check-after-progress 0.20 \
  --timeout 3600

Exit codes:

Code	Meaning
0	Backtest completed normally
1	Runtime error (print full stack trace)
2	Timeout
3	Early stop: drawdown exceeded threshold, backtest deleted

3. Download Results

Fetch the complete backtest statistics (JSON) and all order records (CSV, paginated at 100 per call to bypass QC's limit).

python3 get_results.py <backtest_id> --output-dir ./results/<name>

Output files:

• <backtest_id>_result.json — Full statistics, charts, runtime data
• <backtest_id>_orders.csv — All orders with symbol, direction, quantity, fill price, fees, timestamps

4. Full Pipeline (One Command)

Run the entire submit-compile-monitor-download pipeline:

python3 run_workflow.py \
  --main-file <path_to_strategy.py> \
  --name "<backtest_name>" \
  --max-dd 40

5. List Historical Backtests

python3 list_backtests.py --limit 10

QC API Client

The qc_api/client.py module wraps the QuantConnect REST API v2:

• create_backtest(project_id, name, compile_id) — Start a backtest
• read_backtest(project_id, backtest_id) — Get status and results
• read_backtest_orders(project_id, backtest_id, start, end) — Paginated order fetch
• delete_backtest(project_id, backtest_id) — Delete (used for early-stop)
• update_project_file(project_id, name, content) — Upload code
• compile_project(project_id) — Trigger compilation

Credentials are read from environment variables QC_USER_ID and QC_API_TOKEN.

Rules

• Entry file: Local strategy files are uploaded as main.py. Old Main.py is auto-deleted to prevent conflicts. Do not manually rename files on the QC cloud project.
• One backtest at a time: QC free tier allows only 1 concurrent backtest per project. Never submit a second backtest while one is running — it will silently fail or queue.
• Logs are not available via API: QC REST API v2 does not expose self.log() output. View logs in the QC web UI manually.
• Early-stop threshold: Accepts both fraction (0.40) and percentage (40) formats. Early-stop permanently deletes the backtest — it cannot be recovered.
• Environment variables QC_USER_ID, QC_API_TOKEN, and QC_PROJECT_ID must be set. The skill will raise a clear error if any are missing.
• Do not modify cloud project files outside this skill. The skill assumes it owns main.py and will overwrite any manual changes on the next submission.