Triple Supertrend with EMA and ADX Strategy

Author: ChaoZhang, Date: 2023-09-18 14:02:12
Tags:

Overview

This is a quantitative trading strategy that combines triple supertrend, EMA and ADX indicators. It generates trading signals using a triple supertrend system and applies EMA and ADX as filters to control trade frequency and improve signal quality.

Strategy Logic

  • Use three supertrend systems with different parameters and generate trading signals when all three supertrends agree on direction.

  • Apply EMA as a trend filter, only go long when close is above EMA and go short when close is below EMA.

  • Apply ADX as a trend strength filter, only trade when ADX is above a threshold.

  • Allow re-entry option to adjust profitability and risk control.

Specifically, the long entry condition is when all three supertrends turn bullish, close is above EMA and ADX is higher than the threshold. The short entry condition is when all three supertrends turn bearish, close is below EMA and ADX is higher than the threshold. Exit when any of the supertrends reverse direction.

The strategy also plots supertrend support and resistance lines to aid visual trend determination.

Advantage Analysis

  • Triple supertrend system filters false breakouts and improves entry accuracy.

  • EMA and ADX double filters reduce whipsaw losses and enhance risk management.

  • Re-entry option allows adjusting profitability based on risk preference.

  • Visual supertrend lines help determine trend direction.

Risk Analysis

  • Supertrend and other indicators have lag and may cause late entry or early exit.

  • Too strict filters may miss opportunities.

  • Whipsaws may cause losses in range-bound markets.

  • Allowing re-entry increases trade frequency and slippage costs.

These risks can be reduced by optimizing parameters, filters and using dynamic stops. Position sizing and strict stops should be applied to address uncertain market conditions.

Optimization Directions

This strategy can be optimized in several aspects:

  • Test different parameter combinations to find optimal supertrend and EMA settings.

  • Optimize ADX threshold to reduce false signals.

  • Add other filters like volatility, volume etc.

  • Optimize parameters separately for different products.

  • Build dynamic stop loss mechanisms for better risk control.

  • Explore machine learning for finding better entry and exit rules.

Summary

This strategy utilizes the strengths of triple supertrend systems and augments it with EMA and ADX double filters to effectively improve signal quality and control risks. Further enhancements in parameters, filters, dynamic stops can improve its robustness and adaptiveness. Combined with trend analysis, it provides effective entry and exit signals for quantitative trading.


/*backtest
start: 2023-08-18 00:00:00
end: 2023-09-17 00:00:00
period: 2h
basePeriod: 15m
exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT"}]
*/

// This source code is subject to the terms of the Mozilla Public License 2.0 at https://mozilla.org/MPL/2.0/
// ©kunjandetroja


//@version=5
strategy('Triple Supertrend with EMA and ADX', overlay=true)

m1 = input.float(1,"ATR Multi",minval = 1,maxval= 6,step=0.5,group='ST 1')
m2 = input.float(2,"ATR Multi",minval = 1,maxval= 6,step=0.5,group='ST 2')
m3 = input.float(3,"ATR Multi",minval = 1,maxval= 6,step=0.5,group='ST 3')
p1 = input.int(10,"ATR Multi",minval = 5,maxval= 25,step=1,group='ST 1')
p2 = input.int(15,"ATR Multi",minval = 5,maxval= 25,step=1,group='ST 2')
p3 = input.int(20,"ATR Multi",minval = 5,maxval= 25,step=1,group='ST 3')
len_EMA = input.int(200,"EMA Len",minval = 5,maxval= 250,step=1)
len_ADX = input.int(14,"ADX Len",minval = 1,maxval= 25,step=1)
len_Di = input.int(14,"Di Len",minval = 1,maxval= 25,step=1)
adx_above = input.float(25,"adx filter",minval = 1,maxval= 50,step=0.5)
var bool long_position = false
adx_filter = input.bool(false, "Add Adx & EMA filter")
renetry = input.bool(true, "Allow Reentry")

f_getColor_Resistance(_dir, _color) =>
    _dir == 1 and _dir == _dir[1] ? _color : na
f_getColor_Support(_dir, _color) =>
    _dir == -1 and _dir == _dir[1] ? _color : na

[superTrend1, dir1] = ta.supertrend(m1, p1)
[superTrend2, dir2] = ta.supertrend(m2, p2)
[superTrend3, dir3] = ta.supertrend(m3, p3)
EMA = ta.ema(close, len_EMA)
[diplus,diminus,adx] = ta.dmi(len_Di,len_ADX)

// ADX Filter
adxup = adx > adx_above and close > EMA
adxdown = adx > adx_above and close < EMA

sum_dir = dir1 + dir2 + dir3

dir_long = if(adx_filter == false)
    sum_dir == -3
else
    sum_dir == -3 and adxup
dir_short = if(adx_filter == false)
    sum_dir == 3
else
    sum_dir == 3 and adxdown
Exit_long = dir1 == 1 and dir1 != dir1[1]
Exit_short = dir1 == -1 and dir1 != dir1[1]

// BuySignal = dir_long and dir_long != dir_long[1]
// SellSignal = dir_short and dir_short != dir_short[1]
// if BuySignal
//     label.new(bar_index, low, 'Long', style=label.style_label_up)
// if SellSignal
//     label.new(bar_index, high, 'Short', style=label.style_label_down)

longenter = if(renetry == false)
    dir_long and long_position == false
else
    dir_long
shortenter = if(renetry == false)
    dir_short and long_position == true
else
    dir_short
if longenter
    long_position := true
if shortenter
    long_position := false

strategy.entry('BUY', strategy.long, when=longenter)
strategy.entry('SELL', strategy.short, when=shortenter)   
strategy.close('BUY', Exit_long)
strategy.close('SELL', Exit_short)

buy1 = ta.barssince(dir_long)
sell1 = ta.barssince(dir_short)

colR1 = f_getColor_Resistance(dir1, color.red)
colS1 = f_getColor_Support(dir1, color.green)

colR2 = f_getColor_Resistance(dir2, color.orange)
colS2 = f_getColor_Support(dir2, color.yellow)

colR3 = f_getColor_Resistance(dir3, color.blue)
colS3 = f_getColor_Support(dir3, color.maroon)

plot(superTrend1, 'R1', colR1, linewidth=2)
plot(superTrend1, 'S1', colS1, linewidth=2)

plot(superTrend2, 'R1', colR2, linewidth=2)
plot(superTrend2, 'S1', colS2, linewidth=2)

plot(superTrend3, 'R1', colR3, linewidth=2)
plot(superTrend3, 'S1', colS3, linewidth=2)

// // Intraday only
// var int new_day = na
// var int new_month = na
// var int new_year = na
// var int close_trades_after_time_of_day = na

// if dayofmonth != dayofmonth[1]
//     new_day := dayofmonth
// if month != month[1]
//     new_month := month
// if year != year[1]
//     new_year := year
// close_trades_after_time_of_day := timestamp(new_year,new_month,new_day,15,15)

// strategy.close_all(time > close_trades_after_time_of_day) 


More