0bps

Strategy vault / Machine learning

Auto Trading Strategy Based on Multiple Indicators and Dynamic Stop Loss

This strategy combines multiple indicators like fast, medium, slow MA lines and MACD to generate comprehensive trading signals, and uses dynamic stop loss based on ATR to control risk level. It is suitable for…

Author
ChaoZhang
Complexity
Intermediate · 145 lines
Parameters
30
Backtest bar
1h · intraday
Tested on
BTC/USDT · futures

Overview

This strategy combines multiple indicators like fast, medium, slow MA lines and MACD to generate comprehensive trading signals, and uses dynamic stop loss based on ATR to control risk level. It is suitable for medium-term automated trading.

Strategy Logic

The strategy mainly utilizes EMA, MACD and ATR indicators. EMA fast, medium and slow lines form the trend judgment system. MACD generates trading signals. ATR sets stop loss lines dynamically. Specifically, the trend direction is determined by the combination of the EMA lines. MACD crossover 0 is the entry signal. Enter long when fast line crosses above medium line, and exit when crossing below. Enter short when fast line crosses below medium line, and exit when crossing above. Stop loss adjusts dynamically based on ATR to control risk according to market volatility.

Advantage Analysis

  • Combining multiple indicators makes trading signals accurate and reliable.
  • Fast, medium and slow EMA system provides clear trend judgment.
  • MACD assists entry to avoid false breakout.
  • Dynamic stop loss manages risk better.
  • The mechanical strategy suits automated trading well.

Risks and Improvements

  • Complex parameters require extensive optimization.
  • Complex logic of multiple indicators makes manual trading difficult.
  • Should add other filters like volume to avoid being trapped.
  • Can consider optimizing to machine learning strategy for automated parameter tuning.

Summary

The strategy combines advantages of multiple indicators for accurate trend judgment and drawdown control. Further enhancements like parameter optimization and adding filters can improve robustness. Overall it is a typical and reliable strategy suitable for medium-term automated trading with great practical value.

Parameters

NameDefaultWhat it controls
v_input_17EMA_Smooth_Period
v_input_27ST_VWAP_Period
v_input_3falseVWAP_TUNING_MULT
v_input_413ATR Period
v_input_52ATR Multiplier
v_input_6trueShow Buy/Sell Labels ?
v_input_7trueHighlight State ?
v_input_822StopLoss_Long_Adjust
v_input_916StopLoss_Short_Adjust
v_input_10truefastLength
v_input_114medLength
v_input_1224slowLength
v_input_138signalLength
v_input_142ATR_Signal_Period
v_input_150.986ATR_SIGNAL_FINE_TUNE
v_input_16trueStopLoss_Initial_Short
v_input_175StopLoss_Initial_Long
v_input_1842VOLUME_CHECK_SHORT
v_input_1916VOLUME_CHECK_LONG
v_input_20falseMAX_LOSS
v_input_21falseFrom Minute
v_input_22falseFrom Hour
v_input_23trueFrom Day
v_input_24trueFrom Month
v_input_252019From Year
v_input_26falseTill Minute
v_input_27falseTill Hour
v_input_28trueTill Day
v_input_29trueTill Month
v_input_302021Till Year

Source Pine Script

/*backtest
start: 2023-09-18 00:00:00
end: 2023-09-18 21:00:00
period: 1h
basePeriod: 15m
exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT"}]
*/

//@version=4
strategy("STRAT_STEMWAP", overlay=true, pyramiding = 0, default_qty_value = 10, slippage = 3)

EMA_Smooth_Period = input(7, minval=1)
ST_EMA = ema(close, EMA_Smooth_Period)

ST_VWAP_Period = input(7, minval=1)
VWAP_TUNING_MULT = input(type=input.float, defval=0.000)
ST_VWAP = ema(vwap,ST_VWAP_Period)


ST_VWAP_TUNING = VWAP_TUNING_MULT * (ST_EMA - ST_VWAP)


length = input(title="ATR Period", type=input.integer, defval=13)
mult = input(title="ATR Multiplier", type=input.float, step=0.1, defval=2.0)
showLabels = input(title="Show Buy/Sell Labels ?", type=input.bool, defval=true)
highlightState = input(title="Highlight State ?", type=input.bool, defval=true)

atr = mult * atr(length)


StopLoss_Long_Adjust = input(22.00, type=input.float)
StopLoss_Short_Adjust = input(16.00, type=input.float)


longStop = (ST_EMA) - atr - (ST_VWAP_TUNING) - StopLoss_Long_Adjust
longStopPrev = nz(longStop[1], longStop)
longStop := (close[1]) > longStopPrev ? max(longStop, longStopPrev) : longStop

shortStop = (ST_EMA) + atr - (ST_VWAP_TUNING) + StopLoss_Short_Adjust
shortStopPrev = nz(shortStop[1], shortStop)
shortStop := (close[1]) < shortStopPrev ? min(shortStop, shortStopPrev) : shortStop

dir = 1
dir := nz(dir[1], dir)
dir := dir == -1 and (close) > shortStopPrev ? 1 : dir == 1 and (close) < longStopPrev ? -1 : dir


fastLength = input(1, minval=1), medLength=input(4, minval=1), slowLength=input(24, minval=1), signalLength=input(8,minval=1)
fastMA = ema(close, fastLength)
//,fastMA1 = ema(close[1], fastLength), fastMA2 = ema(close[2], fastLength),fastMA3 = ema(close[3], fastLength),fastMA4 = ema(close[4], fastLength),fastMA5 = ema(close[5], fastLength), fastMA6 = ema(close[6], fastLength), fastMA7 = ema(close[7], fastLength),fastMA8 = ema(close[8], fastLength),fastMA9 = ema(close[9], fastLength),fastMA10 = ema(close[10], fastLength),fastMA11 = ema(close[11], fastLength),fastMA12 = ema(close[12], fastLength),fastMA13 = ema(close[13], fastLength)
medMA = ema(close, medLength)
//, medMA1 = ema(close[1], medLength), medMA2 = ema(close[2], medLength), medMA3 = ema(close[3], medLength), medMA4 = ema(close[4], medLength), medMA5 = ema(close[5], medLength), medMA6 = ema(close[6], medLength), medMA7 = ema(close[7], medLength), medMA8 = ema(close[8], medLength), medMA9 = ema(close[9], medLength), medMA10 = ema(close[10], medLength), medMA11 = ema(close[11], medLength), medMA12 = ema(close[12], medLength), medMA13 = ema(close[13], medLength)
slowMA = ema(close, slowLength)
//, slowMA1 = ema(close[1], slowLength), slowMA2 = ema(close[2], slowLength), slowMA3 = ema(close[3], slowLength), slowMA4 = ema(close[4], slowLength), slowMA5 = ema(close[5], slowLength), slowMA6 = ema(close[6], slowLength), slowMA7 = ema(close[7], slowLength), slowMA8 = ema(close[8], slowLength), slowMA9 = ema(close[9], slowLength), slowMA10 = ema(close[10], slowLength), slowMA11 = ema(close[11], slowLength), slowMA12 = ema(close[12], slowLength), slowMA13 = ema(close[13], slowLength)
macd = fastMA - slowMA
//, macd1 = fastMA1 - slowMA1, macd2 = fastMA2 - slowMA2, macd3 = fastMA3 - slowMA3, macd4 = fastMA4 - slowMA4, macd5 = fastMA5 - slowMA5, macd6 = fastMA6 - slowMA6, macd7 = fastMA7 - slowMA7, macd8 = fastMA8 - slowMA8, macd9 = fastMA9 - slowMA9, macd10 = fastMA10 - slowMA10, macd11 = fastMA11 - slowMA11, macd12 = fastMA12 - slowMA12, macd13 = fastMA13 - slowMA13
fmacd = fastMA - medMA
//, fmacd1 = fastMA1 - medMA1,fmacd2 = fastMA2 - medMA2,fmacd3 = fastMA3 - medMA3,fmacd4 = fastMA4 - medMA4,fmacd5 = fastMA5 - medMA5, fmacd6 = fastMA6 - medMA6, fmacd7 = fastMA7 - medMA7, fmacd8 = fastMA8 - medMA8, fmacd9 = fastMA9 - medMA9, fmacd10 = fastMA10 - medMA10, fmacd11 = fastMA11 - medMA11, fmacd12 = fastMA12 - medMA12, fmacd13 = fastMA13 - medMA13
smacd = slowMA - medMA
//, smacd1 = slowMA1 - medMA1, smacd2 = slowMA2 - medMA2, smacd3 = slowMA3 - medMA3, smacd4 = slowMA4 - medMA4, smacd5 = slowMA5 - medMA5, smacd6 = slowMA6 - medMA6, smacd7 = slowMA7 - medMA7, smacd8 = slowMA8 - medMA8, smacd9 = slowMA9 - medMA9, smacd10 = slowMA10 - medMA10, smacd11 = slowMA11 - medMA11, smacd12 = slowMA12 - medMA12, smacd13 = slowMA13 - medMA13,


signal = ema(macd, signalLength)
//,signal1 = sma(macd1, signalLength),signal2 = sma(macd2, signalLength),signal3 = sma(macd3, signalLength),signal4 = sma(macd4, signalLength),signal5 = sma(macd5, signalLength),signal6 = sma(macd6, signalLength),signal7 = sma(macd7, signalLength),signal8 = sma(macd8, signalLength),signal9 = sma(macd9, signalLength),signal10 = sma(macd10, signalLength),signal11 = sma(macd11, signalLength),signal12 = sma(macd12, signalLength),signal13 = sma(macd13, signalLength),
fsignal = ema(fmacd, signalLength)
//, fsignal1 = sma(fmacd1, signalLength), fsignal2 = sma(fmacd2, signalLength), fsignal3 = sma(fmacd3, signalLength), fsignal4 = sma(fmacd4, signalLength), fsignal5 = sma(fmacd5, signalLength), fsignal6 = sma(fmacd6, signalLength), fsignal7 = sma(fmacd7, signalLength), fsignal8 = sma(fmacd8, signalLength), fsignal9 = sma(fmacd9, signalLength), fsignal10 = sma(fmacd10, signalLength), fsignal11 = sma(fmacd11, signalLength), fsignal12 = sma(fmacd12, signalLength), fsignal13 = sma(fmacd13, signalLength),
ssignal = ema(smacd, signalLength)
//, ssignal1 = sma(smacd1, signalLength), ssignal2 = sma(smacd2, signalLength), ssignal3 = sma(smacd3, signalLength), ssignal4 = sma(smacd4, signalLength), ssignal5 = sma(smacd5, signalLength), ssignal6 = sma(smacd6, signalLength), ssignal7 = sma(smacd7, signalLength), ssignal8 = sma(smacd8, signalLength), ssignal9 = sma(smacd9, signalLength), ssignal10 = sma(smacd10, signalLength), ssignal11 = sma(smacd11, signalLength), ssignal12 = sma(smacd12, signalLength), ssignal13 = sma(smacd13, signalLength),


ATR_Signal_Period = input(2, type=input.integer, minval=1, maxval=2000)


SetStopLossShort = 0.0
SetStopLossShort := if(strategy.position_size < 0)
    StopLossShort = shortStop
    min(StopLossShort,SetStopLossShort[1])


plot(SetStopLossShort, style = plot.style_cross, color = color.yellow)    


SetStopLossLong = 0.0
SetStopLossLong := if(strategy.position_size > 0)
    StopLossLong = longStop
    max(StopLossLong,SetStopLossLong[1])


plot(SetStopLossLong, style = plot.style_cross, color = color.purple)

ATR_SIGNAL_FINE_TUNE = input(0.986, type=input.float)  

tol_atr = atr(ATR_Signal_Period)*ATR_SIGNAL_FINE_TUNE

StopLoss_Initial_Short = input(1.00, type=input.float)
StopLoss_Initial_Long = input(5.00, type=input.float)



VOLUME_CHECK_SHORT = input(42)
VOLUME_CHECK_LONG = input(16)

MAX_LOSS = input(0.00, type=input.float)

//Custom Time Interval
fromMinute = input(defval = 0, title = "From Minute", minval = 0, maxval = 60)
fromHour = input(defval = 0, title = "From Hour", minval = 0, maxval = 24)
fromDay = input(defval = 1, title = "From Day", minval = 1)
fromMonth = input(defval = 1, title = "From Month", minval = 1)
fromYear = input(defval = 2019, title = "From Year", minval = 1900)
tillMinute = input(defval = 0, title = "Till Minute", minval = 0, maxval = 60)
tillHour = input(defval = 0, title = "Till Hour", minval = 0, maxval = 24)
tillDay = input(defval = 1, title = "Till Day", minval = 1)
tillMonth = input(defval = 1, title = "Till Month", minval = 1)
tillYear = input(defval = 2021, title = "Till Year", minval = 1900)
timestampStart = timestamp(fromYear,fromMonth,fromDay,fromHour,fromMinute)
timestampEnd = timestamp(tillYear,tillMonth,tillDay,tillHour,tillMinute)



if ( strategy.position_size <= 0 and ((fsignal[1] -fsignal) <= 0) and volume > VOLUME_CHECK_LONG and ( cross(signal, macd) or cross(signal[1], macd[1]) or cross(signal[2], macd[2]) or cross(signal[3], macd[3]) or cross(signal[4], macd[4]) or cross(signal[5], macd[5]) or cross(signal[6], macd[6]) or cross(signal[7], macd[7]) or  cross(signal[8], macd[8]) or  cross(signal[9], macd[9]) or  cross(signal[10], macd[10]) or  cross(signal[11], macd[11]) or  cross(signal[12], macd[12]) or  cross(signal[13], macd[13])  or cross(fmacd, macd) or cross(fmacd[1],macd[1]) or  cross(fmacd[2],macd[2]) or cross(fmacd[3],macd[3]) or cross(fmacd[4],macd[4])or cross(fsignal, fmacd)  or cross(fmacd, smacd) )  and (  (crossover(close,open+tol_atr) or crossover(close[1],open[1]+tol_atr[1]) or crossover(close[2],open[2]+tol_atr[2]) or crossover(close[3],open[3]+tol_atr[3]) or crossover(close[4],open[4]+tol_atr[4]) or crossover(close[5],open[5]+tol_atr[5]) or crossover(close[6],open[6]+tol_atr[6]) or crossover(close[7],open[7]+tol_atr[7]) or crossover(close[8],open[8]+tol_atr[8]) or crossover(close[9],open[9]+tol_atr[9]) or crossover(close[10],open[10]+tol_atr[10]) ) or ( (cross(ssignal, smacd) or cross(ssignal[1],smacd[1]) or cross(ssignal[2],smacd[2])) and ( (ssignal - ssignal[1]) > 0 ) ) )  )
    strategy.exit("SELL")
    strategy.entry("BUY", strategy.long)
    strategy.exit("BUY_STOP","BUY", stop = close - StopLoss_Initial_Long)
    

if ( (dir == -1 and dir[1] == 1 and dir[2] == 1 and dir[3] == 1 and dir[4] == 1 ) and strategy.position_size >= 0 and volume > VOLUME_CHECK_SHORT and ((fsignal[1] -fsignal) >= 0)  and (  crossunder(close, open - tol_atr) or crossunder(close[1], open[1] - tol_atr[1]) or crossunder(close[2], open[2] - tol_atr[2]) or crossunder(close[3], open[3] - tol_atr[3]) )  and ( cross(signal, macd) or cross(signal[1], macd[1]) or cross(signal[2], macd[2]) or cross(signal[3], macd[3]) or cross(signal[4], macd[4]) or cross(signal[5], macd[5]) or cross(signal[6], macd[6]) or cross(signal[7], macd[7]) or  cross(signal[8], macd[8]) or  cross(signal[9], macd[9]) or  cross(signal[10], macd[10]) or  cross(signal[11], macd[11])  or  cross(signal[12], macd[12]) or  cross(signal[13], macd[13])  )  )
    strategy.exit( "BUY")
    strategy.entry("SELL", strategy.short)
    strategy.exit("SELL_STOP","SELL", stop = close + StopLoss_Initial_Short)
    

strategy.close_all(when = strategy.openprofit <  (-1 * MAX_LOSS) )


// if (strategy.max_contracts_held_long > 0 )
//     strategy.exit("BUY_TRAIL_STOP","BUY", stop = longStop - StopLoss_Long_Adjust)
    
// if (strategy.max_contracts_held_short > 0 )    
//     strategy.exit("SELL_TRAIL_STOP","SELL", stop = shortStop + StopLoss_Short_Adjust)


    
    //strategy.exit("BUY_TRAIL_STOP","BUY", stop = SetStopLossLong)
    //strategy.exit("SELL_TRAIL_STOP","SELL", stop = SetStopLossShort)     

Published by ChaoZhang on the FMZ strategy square · last modified 2023-09-26. Collected in the Quant Trading Vault. Unvalidated research: re-test with your own data and costs before trusting it.

Build this with AgenKit

A spec is not a trading system. Paste this into your coding agent and AgenKit runs a specialist team — architecture, test-first build, risk gates, review — with paper trading on by default.

in Claude Code, Cursor, Codex & 4 more
/agenkit build a production-grade trading bot with a backtest engine, risk controls and paper trading by default from strategies/Auto-Trading-Strategy-Based-on-Multiple-Indicators-and-Dynamic-Stop-Loss.md
  1. npx agenkit install engineering-kit in your project
  2. Paste the command into your agent and approve each phase
  3. Get tests, backtest, risk controls and a paper-trading switch

Understand it first

Before you run this, read the lesson behind the idea and how to backtest it honestly. Specs like this one are the survivors of an unknown number of trials; deflate their results accordingly.

Related specs

esc

Try , , or .

↑↓ move↵ open