Momentum breakout system
Trading · sample build
A multi-timeframe breakout strategy for index futures with volatility-scaled position sizing.
The problem
A trader had a discretionary breakout method that worked in strong trends and bled money in choppy markets. They wanted it written down, tested and automated.
The approach
Turned the method into explicit rules, added a volatility-regime filter to stand aside in chop, and sized positions by recent ATR. Tested walk-forward, then paper traded for a month against the backtest before going live.
How it was tested
Rules were frozen before the last two years of data were opened. Walk-forward runs re-optimised every six months on the preceding three years, and the breakout length was taken from the middle of a stable range rather than its best value. Costs were doubled as a stress test, and a month of paper trading was compared trade by trade with what the backtest did over the same days.
What was handed over
Strategy code, walk-forward test suite, paper-trading comparison report, deployment on the client’s server with daily loss limits and an alert channel.
Stack
- Python
- pandas
- NumPy
- Interactive Brokers API
- PostgreSQL
- AWS