This strategy is based on the Dynamic Moving Average indicator, combined with Bollinger Bands and RSI for trade signal filtering. It implements a trend following long only strategy. The strategy judges the trend by calculating the change of Heiken Ashi closing price Dynamic Moving Average and compares it with Bollinger Bands to generate trade signals. With the RSI filter, it can effectively identify trend explosive points for trend tracking.
The core of this strategy is to calculate the change of Heiken Ashi closing price Dynamic Moving Average. Specifically, it calculates the difference between the current bar’s MA and the MA of previous two bars, then multiplies it by a sensitivity coefficient to get the accurate MA change value.
Then this change value is compared with the difference between Bollinger Bands upper band and lower band. If the MA change is greater than the BB difference, it is considered as a “trend explosion”. When the explosion is positive, i.e. the MA change is positive, it generates a long signal and green bar. When the explosion is negative, i.e. the MA change is negative, it generates a close signal and red bar.
In addition, this strategy has a RSI filter that only allows long signals when RSI is higher than a threshold, avoiding the risk of trend reversal.
Risk control methods include: proper parameter tuning for robustness, combining other indicators to judge trend reversal, use only in clear long-term trends, etc.
There is some room for further optimization:
Try different price sources like close, moving averages etc. for better smoothing
Adjust MA and BB period parameters for optimization across different products
Try ratio relationship instead of sensitivity coefficient for more intuitive indicator value
Add other filters like trendlines, volume etc. to improve signal quality
Develop short strategy based on indicator patterns
Incorporate stop loss mechanisms for better risk control
Overall this is a relatively stable trend following strategy. It uses dynamic moving average to determine trend direction, BB to identify explosive points, RSI to filter false signals, realizing a long only trend system. But it also has some risks, requiring parameter tuning for different products and timeframes, and inability to profit from downtrends. There is room for further improvements like enhancing signal quality, developing short strategy, adding stop loss etc. to achieve better performance.
/*backtest start: 2022-11-08 00:00:00 end: 2023-11-14 00:00:00 period: 1d basePeriod: 1h exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT"}] */ //@version=5 ///////////Original Script Courtesy of Lazy_Bear.... Absolute Legend\\\\\\\\\\\\\\\ strategy('SmoothedWaddah', overlay=false, initial_capital=1) sensitivity = input(150, title='Sensitivity') fastLength = input(20, title='MacD FastEMA Length') slowLength = input(40, title='MacD SlowEMA Length') channelLength = input(20, title='BB Channel Length') mult = input(1.5, title='BB Stdev Multiplier') RSI14filter = input(40, title='RSI Value trade filter') ////////////MacD Calculation of price////////////////////////////// calc_macd(source, fastLength, slowLength) => fastMA = ta.ema(source, fastLength) slowMA = ta.ema(source, slowLength) fastMA - slowMA /////////BolingerBand Calculation of Price/////////////////////// calc_BBUpper(source, length, mult) => basis = ta.sma(source, length) dev = mult * ta.stdev(source, length) basis + dev calc_BBLower(source, length, mult) => basis = ta.sma(source, length) dev = mult * ta.stdev(source, length) basis - dev //////heinkenashi chart call for closing price "smoothing mechanism"\\\\\\\\\\\\\\\\\\\\\\\\\\\ point = request.security(ticker.heikinashi(syminfo.tickerid), timeframe.period, close) ////////////////////T1 is change in MacD current candle from previous candle Sensitivy amplifies calculation///////////////////// t1 = (calc_macd(point, fastLength, slowLength) - calc_macd(point[1], fastLength, slowLength)) * sensitivity //////////////////////T2 is T1 from two candles prior\\\\\\\\\\\\\\\\\\\\\\\\\\\ t2 = (calc_macd(point[2], fastLength, slowLength) - calc_macd(point[3], fastLength, slowLength)) * sensitivity ////////////////E1 is difference in bolinger band upper and lower...E2 is E1 from one candle prior not needed////////////// e1 = calc_BBUpper(ohlc4, channelLength, mult) - calc_BBLower(ohlc4, channelLength, mult) //e2 = (calc_BBUpper(close[1], channelLength, mult) - calc_BBLower(close[1], channelLength, mult)) //////signal bar printing.. Up if MacD positive .. Down if MacD negative////////// trendUp = t1 >= 0 ? t1 : 0 trendDown = t1 < 0 ? -1 * t1 : 0 ///////plots difference in macD*Sensitivity, color change if increasing or decreasing. //////color is green/lime if explosion is up \ color is red/orange if explosion is down///////// plot(trendUp, style=plot.style_columns, linewidth=1, color=trendUp < trendUp[1] ? color.new(color.lime,45) : color.new(color.green,45), title='UpTrend') plot(trendDown, style=plot.style_columns, linewidth=1, color=trendDown < trendDown[1] ? color.new(color.orange,45) : color.new(color.red,45), title='DownTrend') plot(e1, style=plot.style_line, linewidth=2, color=color.new(#A0522D, 0), title='ExplosionLine') ////////////Entry conditions and Concept///////////////////// ////////////Long Only System. T1 is measuring the distance between MACD EMA's. This is Multiplied ////////////by the sensitivity so that it can be compared to the difference between BollingerBand. /////////////{this could have been a ratio maybe i will work with that in a different script.} /////////////I found that 135-175 sensitivy allows for values to be compared on most charts..... ////////////If the (difference between the EMA)*(Sensitivity) is greater than (BB upper line- BB lower line) ////////////it is considered an explosion in either the downside or the upside.The indicator will print ///////////a bar higher than the trigger line either green or red (up or down respectively)////////////////// longCondition = trendUp > e1 and ta.rsi(close, 14) > RSI14filter if longCondition strategy.entry('up', strategy.long) strategy.close('up', trendDown > e1)