- Prefer fewer conditions. A rule with seven filters can look perfect on one backtest and never trigger again. Two or three clear conditions are easier to trust.
- Test more than one timeframe. The default is 15-minute candles, which gives a thousand-candle backtest about ten days of data. Try 1h for a longer view.
- Always set a way out. Pair your exit rule with a take-profit and a stop-loss so a position never waits forever.
- Run it on paper first. A backtest shows the past. A few days on paper shows whether the rule still behaves on fresh candles.
Signal bots are long-only and spot-only: they buy coins with your own funds and sell them back. There is no shorting and no leverage.