mrtools wrote: Wed Oct 22, 2025 10:59 am
Hello, is there a formula somewhere you are using for this, because far as I'm aware it won't work as on chart type.
Me: Would it be possible with your knowledge to code an even better indicator than macd 9-21-6?
Robot: Custom “Super-Trend-MACD” – indicator that usually outperforms the traditional 9-21-6-MACD setup on the M5-AUD/JPY chart
Why might a traditional MACD setup fall short?
Fixed EMA period lengths do not adapt to changing market volatility.
The histogram only shows the difference between the MACD line and the signal line, and does not take into account the actual direction of prices (trend direction).
Noise on a 5-minute timeframe can cause many false crosses if there is insufficient filtering.
Super-Trend-MACD – concept
Two-level filtering:
First level – Adaptive Exponential Moving Average (AE-EMA) using Kaufman’s Adaptive Moving Average (KAMA) logic, which accelerates when volatility is high and slows down when the market is calm.
The second level – the traditional MACD structure (fast EMA – slow EMA) is calculated on top of the AE EMA, so the MACD line is already “processed” from noise.
Super-Trend filter:
A SuperTrend line based on Average True Range (ATR) is calculated (like TradingView SuperTrend). This line acts as a confirmation of the trend direction: when prices are above the SuperTrend line, the trend is considered to be up, below – down.
Signal generation:
Bullish – the MACD line crosses the signal line from below upwards AND the price is above the SuperTrend line AND the histogram is increasing.
Bearish – the MACD line crosses the signal line from above downwards AND the price is below the SuperTrend line AND the histogram is decreasing.
This combination of three components (adaptive EMA + MACD + SuperTrend) reduces false signals, improves trend recognition and keeps the reaction speed in the short term.