I Coded a Supertrend Strategy Backtest - and Here Are the Results
No better than luck
Small profit, too noisy to call an edge
- trades tested
- 774
- per trade after costs
- +2.77bp
- winners
- 32%
- years of data
- 4.8
Multi-timeframe Supertrend (10,3): on the 30m chart, go long when the 30m Supertrend turns green while the 4h Supertrend is green, short when it turns red while the 4h is red. No take profit; the 30m Supertrend line is the trailing stop.
The exact rules we tested
- 01Instrument: BTCUSDT | Signal timeframe: 30m | Direction: both
- 02LONG when: st.dir crosses above 0
- 03SHORT when: st.dir crosses below 0
- 04Higher-timeframe inputs (closed-bar values only, no repaint): sth on 4h
- 05Long filter: sth.dir > 0
- 06Short filter: sth.dir < 0
- 07Entry: market at next bar open
- 08Stop: at the st.line (longs) / st.line (shorts) level
- 09Exit longs when: st.dir crosses below 0
- 10Exit shorts when: st.dir crosses above 0