These tips help you build cleaner strategies that backtest better and avoid common logic problems. Whether you're using a template or the custom builder, following these guidelines will save you time and produce more reliable results.
⚡ Best Practices
- Keep entry conditions clear and specific. Vague descriptions like "when price looks bullish" can't be coded reliably. Instead, describe exact conditions like "when price closes above the 20 EMA."
- Avoid contradictory confirmations. Selecting both bullish-only and bearish-only confirmations together can create logic that never triggers. Make sure your conditions work in the same direction.
- Don't overfilter with too many rules. Each additional confluence reduces the number of trades. Start with 2-3 key conditions and add more only if backtests show improvement.
- Always define stop loss and take profit. A strategy without exit logic will hold positions indefinitely. Make sure both SL and TP are selected and clearly defined.
- Use time and session filters carefully. Restricting trades to a narrow window reduces sample size. Make sure your chosen session actually has enough activity for your instrument.
- Avoid repainting or future-looking conditions. Indicators that change historical values (like repainting pivots) produce misleading backtests. Use confirmed, closed-bar data only.
- Keep higher timeframe logic consistent. If your key level is on the 15m chart, don't expect it to update on every 1m bar. Higher TF signals only change when that TF bar closes.
- Make rules objective, not vague. "Strong momentum" is subjective. "RSI above 70" or "displacement candle with body > 2x ATR" is objective and codeable.
- Test one change at a time. If you change three things and results improve, you won't know which change helped. Iterate with single modifications.
- Make sure every confluence can actually trigger. Some combinations are theoretically possible but practically never align. If your backtest shows zero trades, simplify your conditions.
⚠️ Common Mistakes
- Too many confirmations causing zero trades. Requiring 5+ confluences to all align at the same time is extremely rare. Most real strategies use 2-3 conditions.
- Conflicting bullish and bearish logic. For example, requiring price to be in a premium zone (bearish) while also looking for bullish OB entries creates a contradiction.
- Undefined or missing exits. Entering trades without a clear SL/TP means the strategy can't properly close positions. Always define both before generating.
- Timeframe-dependent logic bugs. A condition that works on 5m may never trigger on 1h because the bar structure is different. Match your timeframes to your strategy's logic.
- Conditions based on visual ideas that aren't codeable. "I can see a head and shoulders" is pattern recognition that's hard to express in exact rules. Stick to indicator-based or price-action conditions with clear math.
