Dual directional ADX Trading Strategy

Pine ScriptMachine learning Trend following SMA ADX / DMI

Author: ChaoZhang
Complexity: Intermediate · 198 lines
Parameters: 12
Backtest bar: 1h · intraday
Tested on: BTC/USDT · futures

Jump to code View in vault Original on FMZ

Overview

The dual-directional ADX trading strategy is a quantitative strategy that implements dual-directional trading using the Average Directional Index (ADX) indicator. The strategy generates trading signals by calculating the difference between the ADX indicator and the DIPlus and DIMinus indicators and setting thresholds to determine long and short entries for profit.

Strategy Logic

Calculate the True Range
Calculate the Directional Movement Plus and Directional Movement Minus
Calculate the Smoothed True Range
Calculate the Smoothed Directional Movement Plus and Smoothed Directional Movement Minus
Calculate the DIPlus, DIMinus and ADX indicators
Calculate the difference between DIPlus & ADX and DIMinus & ADX
Set thresholds for long and short trade differences
Generate trading signals when difference exceeds thresholds
Create buy and sell orders

The core of this strategy is using ADX and directional movement indicators to determine trend direction and strength, combined with difference threshold rules to filter signals and automate trading.

Advantage Analysis

ADX accurately captures market trend
Difference threshold rules effectively filter false signals
Dual-directional trading fully captures long and short opportunities
Fully automated trading without manual intervention
Clear strategy logic, easy to understand and modify

Risk Analysis

ADX has lag, may miss trend turning points
Increased risk from dual-directional trading, larger losses
Improper parameter setting may cause over-trading
Backtest data cannot represent real market, real trading risk exists

Solutions:

Confirm signals with other indicators
Optimize parameters, control trade frequency
Strict position sizing to manage position size

Optimization Directions

Optimize ADX parameters to improve sensitivity
Add other indicators to filter signals
Apply machine learning to optimize parameters
Use advanced stop loss strategies to control losses
Combine with model predictions for more accurate signals

Conclusion

The dual-direction ADX trading strategy overall is a very practical quantitative strategy. It identifies trends using the ADX indicator and captures trading opportunities in both directions. Meanwhile, it uses difference thresholds to validate signal effectiveness. The strategy has clear and simple logic that is easy to modify and optimize. It is a dual-directional trend following system. Further improvements in stability and profitability can be achieved through parameter optimization, stop loss strategies, and signal filtration.

Parameters

Name	Default	What it controls
v_input_1	14	len
v_input_2	20	th
v_input_3	10	Long Difference
v_input_4	10	Short Difference
v_input_bool_1	true	(?Monthly Performance)Show Monthly Monthly Performance ?
v_input_5	0	Location: Bottom Right \| Top Right \| Top Left \| Bottom Left \| Middle Right \| Bottom Center
v_input_6	0	Size: Small \| Tiny \| Normal \| Large
v_input_color_1	#07e2f2	Background Color
v_input_color_2	#000000	Month/Year Heading Color
v_input_color_3	white	Month PnL Data Color
v_input_color_4	#000000	Year PnL Data Color
v_input_color_5	white	Table Border Color

Source Pine Script

/*backtest
start: 2023-12-01 00:00:00
end: 2023-12-31 23:59:59
period: 1h
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/
// © MAURYA_ALGO_TRADER

//@version=5
strategy("Monthly Performance", overlay=true)


len = input(14)
th = input(20)

TrueRange = math.max(math.max(high - low, math.abs(high - nz(close[1]))), math.abs(low - nz(close[1])))
DirectionalMovementPlus = high - nz(high[1]) > nz(low[1]) - low ? math.max(high - nz(high[1]), 0) : 0
DirectionalMovementMinus = nz(low[1]) - low > high - nz(high[1]) ? math.max(nz(low[1]) - low, 0) : 0

SmoothedTrueRange = 0.0
SmoothedTrueRange := nz(SmoothedTrueRange[1]) - nz(SmoothedTrueRange[1]) / len + TrueRange

SmoothedDirectionalMovementPlus = 0.0
SmoothedDirectionalMovementPlus := nz(SmoothedDirectionalMovementPlus[1]) - nz(SmoothedDirectionalMovementPlus[1]) / len + DirectionalMovementPlus

SmoothedDirectionalMovementMinus = 0.0
SmoothedDirectionalMovementMinus := nz(SmoothedDirectionalMovementMinus[1]) - nz(SmoothedDirectionalMovementMinus[1]) / len + DirectionalMovementMinus

DIPlus = SmoothedDirectionalMovementPlus / SmoothedTrueRange * 100
DIMinus = SmoothedDirectionalMovementMinus / SmoothedTrueRange * 100
DX = math.abs(DIPlus - DIMinus) / (DIPlus + DIMinus) * 100
ADX = ta.sma(DX, len)

// plot(DIPlus, color=color.new(color.green, 0), title='DI+')
// plot(DIMinus, color=color.new(color.red, 0), title='DI-')
// plot(ADX, color=color.new(color.white, 0), title='ADX')
// hline(th, color=color.black)


//diff_1 = math.abs(DIPlus - DIMinus)
diff_2 = math.abs(DIPlus-ADX)
diff_3 = math.abs(DIMinus - ADX)

long_diff = input(10, "Long Difference")
short_diff = input(10, "Short Difference")

buy_condition = diff_2 >=long_diff and diff_3 >=long_diff and (ADX < DIPlus and ADX > DIMinus)
sell_condition = diff_2 >=short_diff and diff_3 >=short_diff and (ADX > DIPlus and ADX < DIMinus)


if buy_condition
    strategy.entry("Long Entry", strategy.long, comment = "Long")
if sell_condition
    strategy.entry("Short Entry", strategy.short, comment = "Short")



// Copy below code to end of the desired strategy script
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
//                                 monthly pnl performance  by Dr. Maurya @MAURYA_ALGO_TRADER                        //
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
show_performance = input.bool(true, 'Show Monthly Monthly Performance ?', group='Monthly Performance')

dash_loc_mp = input("Bottom Right","Location"  ,options=["Top Right","Bottom Right","Top Left","Bottom Left", "Middle Right","Bottom Center"]  ,group='Monthly Performance', inline = "performance")

text_size_mp = input('Small',"Size"  ,options=["Tiny","Small","Normal","Large"]  ,group='Monthly Performance', inline = "performance")

bg_c = input.color( color.rgb(7, 226, 242, 38), "Background Color", group='Monthly Performance')

text_head_color = input.color( color.rgb(0,0,0), "Month/Year Heading Color", group='Monthly Performance')

tab_month_c = input.color( color.white, "Month PnL Data Color", group='Monthly Performance')

tab_year_c = input.color( color.rgb(0,0,0), "Year PnL Data Color", group='Monthly Performance')

border_c = input.color( color.white, "Table Border Color", group='Monthly Performance')



var table_position_mp = dash_loc_mp == 'Top Left' ? position.top_left :
  dash_loc_mp == 'Bottom Left' ? position.bottom_left :
  dash_loc_mp == 'Middle Right' ? position.middle_right :
  dash_loc_mp == 'Bottom Center' ? position.bottom_center :
  dash_loc_mp == 'Top Right' ? position.top_right : position.bottom_right
  
var table_text_size_mp = text_size_mp == 'Tiny' ? size.tiny :
  text_size_mp == 'Small' ? size.small :
  text_size_mp == 'Normal' ? size.normal : size.large

/////////////////

strategy.initial_capital = 50000

/////////////////////////////////////////////

// var bool new_month = na
new_month = ta.change(month) //> 0 ? true : false
newest_month = new_month and strategy.closedtrades >= 1

// profit
only_profit = strategy.netprofit
initial_balance = strategy.initial_capital

// month number
var int month_number = na
month_number := (ta.valuewhen(newest_month, month(time), 0)) //and month(time) > 1 ? (ta.valuewhen(newest_month, month(time), 0) - 1) :  12 //1 to 12

//month_year
var int month_time = na
month_time := ta.valuewhen(newest_month, time, 0) - 2419200000 


var int m_counter = 0
if newest_month
    m_counter += 1



// current month values
var bool new_year = na
new_year := ta.change(year)
curr_m_pnl = only_profit - nz(ta.valuewhen(newest_month, only_profit, 0), 0)
curr_m_number = newest_month ? ta.valuewhen(newest_month, month(time), 0) : month(time)
curr_y_pnl = (only_profit - nz(ta.valuewhen(new_year, only_profit, 0),0)) 



var float [] net_profit_array = array.new_float()
var int [] month_array = array.new_int()
var int [] month_time_array = array.new_int()


if newest_month
    array.push(net_profit_array, only_profit)
    array.push(month_array, month_number)
    array.push(month_time_array, month_time)



var float [] y_pnl_array = array.new_float()
var int [] y_number_array = array.new_int()
var int [] y_time_array = array.new_int()

newest_year = ta.change(year) and strategy.closedtrades >= 1
get_yearly_pnl = nz(ta.valuewhen(newest_year, strategy.netprofit, 0) - nz(ta.valuewhen(newest_year, strategy.netprofit, 1), 0), 0)
get_m_year = ta.valuewhen(newest_year, year(time), 1)
get_y_time = ta.valuewhen(newest_year, time, 0)

if newest_year
    array.push(y_pnl_array, get_yearly_pnl)
    array.push(y_number_array, get_m_year)
    array.push(y_time_array, get_y_time)
var float monthly_profit = na
var int column_month_number = na
var int row_month_time = na

 


var testTable = table.new(position = table_position_mp, columns = 14, rows = 40, bgcolor = bg_c, border_color = border_c, border_width = 1)
if barstate.islastconfirmedhistory and show_performance
    table.cell(table_id = testTable, column = 0, row = 0, text = "YEAR", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 1, row = 0, text = "JAN", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 2, row = 0, text = "FEB", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 3, row = 0, text = "MAR", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 4, row = 0, text = "APR", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 5, row = 0, text = "MAY", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 6, row = 0, text = "JUN", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 7, row = 0, text = "JUL", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 8, row = 0, text = "AUG", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 9, row = 0, text = "SEP", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 10, row = 0, text = "OCT", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 11, row = 0, text = "NOV", text_color = text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 12, row = 0, text = "DEC", text_color =text_head_color, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 13, row = 0, text = "YEAR P/L", text_color = text_head_color, text_size=table_text_size_mp)

    for i = 0 to (array.size(y_number_array) == 0 ? na : array.size(y_number_array) - 1)
        row_y = year(array.get(y_time_array, i)) - year(array.get(y_time_array, 0)) + 1
        table.cell(table_id = testTable, column = 13, row = row_y, text = str.tostring(array.get(y_pnl_array , i), "##.##") + '\n' + '(' + str.tostring(array.get(y_pnl_array , i)*100/initial_balance, "##.##") + ' %)', bgcolor = array.get(y_pnl_array , i) > 0 ? color.green : array.get(y_pnl_array , i) < 0 ? color.red : color.gray, text_color = tab_year_c, text_size=table_text_size_mp)
    curr_row_y = array.size(month_time_array) == 0 ? 1 : (year(array.get(month_time_array, array.size(month_time_array) - 1))) - (year(array.get(month_time_array, 0))) + 1
    table.cell(table_id = testTable, column = 13, row = curr_row_y, text = str.tostring(curr_y_pnl, "##.##") + '\n' + '(' + str.tostring(curr_y_pnl*100/initial_balance, "##.##") + ' %)', bgcolor = curr_y_pnl > 0 ? color.green : curr_y_pnl < 0 ? color.red : color.gray, text_color = tab_year_c, text_size=table_text_size_mp)
    

    for i = 0 to (array.size(net_profit_array) == 0 ? na : array.size(net_profit_array) - 1)
        monthly_profit := i > 0 ? ( array.get(net_profit_array, i) - array.get(net_profit_array, i - 1) ) : array.get(net_profit_array, i) 
        column_month_number := month(array.get(month_time_array, i)) 
        row_month_time :=((year(array.get(month_time_array, i))) - year(array.get(month_time_array, 0)) ) + 1 
        table.cell(table_id = testTable, column = column_month_number, row = row_month_time, text = str.tostring(monthly_profit, "##.##") + '\n' + '(' + str.tostring(monthly_profit*100/initial_balance, "##.##") + ' %)', bgcolor = monthly_profit > 0 ? color.green : monthly_profit < 0 ? color.red : color.gray, text_color = tab_month_c, text_size=table_text_size_mp)
        table.cell(table_id = testTable, column = 0, row =row_month_time, text = str.tostring(year(array.get(month_time_array, i)), "##.##"), text_color = text_head_color, text_size=table_text_size_mp)
       
    curr_row_m = array.size(month_time_array) == 0 ? 1 : (year(array.get(month_time_array, array.size(month_time_array) - 1))) - (year(array.get(month_time_array, 0))) + 1
    table.cell(table_id = testTable, column = curr_m_number, row = curr_row_m, text = str.tostring(curr_m_pnl, "##.##") + '\n' + '(' + str.tostring(curr_m_pnl*100/initial_balance, "##.##") + ' %)', bgcolor = curr_m_pnl > 0 ? color.green : curr_m_pnl < 0 ? color.red : color.gray, text_color = tab_month_c, text_size=table_text_size_mp)
    table.cell(table_id = testTable, column = 0, row =curr_row_m, text = str.tostring(year(time), "##.##"), text_color = text_head_color, text_size=table_text_size_mp)

//============================================================================================================================================================================

Published by ChaoZhang on the FMZ strategy square · last modified 2024-01-30. 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/Dual-directional-ADX-Trading-Strategy.md

npx agenkit install engineering-kit in your project
Paste the command into your agent and approve each phase
Get tests, backtest, risk controls and a paper-trading switch

Get AgenKit How /agenkit works

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.