Quantitative Trend Tracking Strategy

Pine ScriptTrend following Machine learning Breakout RSI MACD EMA SMA Bollinger Bands ADX / DMI

Author: ChaoZhang
Complexity: Intermediate · 133 lines
Parameters: 13
Backtest bar: 5h · swing
Tested on: BTC/USDT · futures

Jump to code View in vault Original on FMZ

Overview

This strategy adopts multiple indicators such as Bollinger Bands, RSI, ADX, MACD to judge market trends and has strong trend identification capability. It takes trend following strategy when indicator signals are bullish simultaneously. It closes positions to stop loss when indicator signals are bearish simultaneously.

Strategy Principle

Use Bollinger Bands to judge whether price is near upper or lower rail to determine whether a trend is formed
Combine with RSI indicator to avoid overbought and oversold areas to prevent false breakouts
Use ADX to determine trend strength and only signal when trend strength is strong
Adopt MACD to judge consistency of short-term and long-term trends
Limit to trading sessions to avoid overnight risks

Through the combination judgement of multiple indicators, it can accurately identify price trends and timely track them when trend occurs to achieve excess returns.

Advantage Analysis

The biggest advantage of this strategy is that the indicator combination judgement is more comprehensive and accurate, which can effectively identify price trends and avoid false signals caused by single indicators.

Specifically, the advantages are:

Bollinger Bands can determine price fluctuation range and strength
RSI avoids buying in overbought areas and selling in oversold areas
ADX determines trend strength, only follows strong trends
MACD judges consistency of short and long term
Limits trading sessions to avoid overnight risks

Through indicator combination judgement, it can maximize reduce false signals and enhance stability of the strategy.

Risk Analysis

The main risks of this strategy come from:

Market events causing indicator failure
Frequent false signals in range-bound markets

For risk 1, reliance on multiple indicators can mitigate the problem of single indicator failure to some extent, but risk management mechanisms still need to be improved.

For risk 2, parameters can be adjusted appropriately to narrow trading range and reduce trading frequency to mitigate risks.

Optimization Directions

The main optimizable aspects of this strategy include:

Add stop loss mechanisms such as trailing stop loss, time stop loss, breakout stop loss, etc. to avoid pulling back too deep
Optimize parameters, adjust indicator parameters combination to find optimum
Add filters such as volume filter to avoid false breakouts in low volume
Incorporate more indicators like KDJ, OBV to improve signal accuracy
Adopt machine learning methods to auto-optimize parameters

By continuous optimization, continuously improve parameter robustness and reduce probabilities of false signals.

Summary

Overall this strategy has relatively strong capability of identifying trend signals through indicator combination judgements which can effectively identify price trends.

But it also has certain risks, risk management and parameter optimization needs to be continuously improved for steady long-term operations. If methods like machine learning can be introduced later to achieve auto-optimization of parameters, it will greatly enhance robustness and profitability of the strategy.

Parameters

Name	Default	What it controls
v_input_1	0930-1500	Trading Session
v_input_2_high	0	Bollinger Band Source: high \| close \| low \| open \| hl2 \| hlc3 \| hlcc4 \| ohlc4
v_input_3	3	Bollinger Band Length
v_input_4	4.989	Bollinger Band Std Dev
v_input_5_close	0	RSI Source: close \| high \| low \| open \| hl2 \| hlc3 \| hlcc4 \| ohlc4
v_input_6	16	RSI Length
v_input_7	39.2	RSI Comparator
v_input_8	14	ADX Length
v_input_9	14	ADX Comparator
v_input_10_high	0	MACD Source: high \| close \| low \| open \| hl2 \| hlc3 \| hlcc4 \| ohlc4
v_input_11	12	MACD Fast
v_input_12	20	MACD Slow
v_input_13	15	MACD Signal

Source Pine Script

/*backtest
start: 2024-01-01 00:00:00
end: 2024-01-31 00:00:00
period: 5h
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/
// © abilash.s.90


dIMinusCalc(adxLen) =>
    
    smoothedTrueRange = 0.0
    smoothedDirectionalMovementMinus = 0.0
    dIMinus = 0.0
    trueRange = 0.0
    directionalMovementMinus = 0.0
    
    trueRange := max(max(high-low, abs(high-nz(close[1]))), abs(low-nz(close[1])))
    directionalMovementMinus := nz(low[1])-low > high-nz(high[1]) ? max(nz(low[1])-low, 0): 0
    
    smoothedTrueRange := nz(smoothedTrueRange[1]) - (nz(smoothedTrueRange[1])/adxLen) + trueRange
    smoothedDirectionalMovementMinus := nz(smoothedDirectionalMovementMinus[1]) - (nz(smoothedDirectionalMovementMinus[1])/adxLen) + directionalMovementMinus
    
    dIMinus := smoothedDirectionalMovementMinus / smoothedTrueRange * 100
    
    dIMinus

dIPlusCalc(adxLen) =>
    
    smoothedTrueRange = 0.0
    smoothedDirectionalMovementPlus = 0.0
    dIPlus =  0.0
    trueRange = 0.0
    directionalMovementPlus = 0.0
    
    trueRange := max(max(high-low, abs(high-nz(close[1]))), abs(low-nz(close[1])))
    directionalMovementPlus := high-nz(high[1]) > nz(low[1])-low ? max(high-nz(high[1]), 0): 0
    
    smoothedTrueRange := nz(smoothedTrueRange[1]) - (nz(smoothedTrueRange[1])/adxLen) + trueRange
    smoothedDirectionalMovementPlus := nz(smoothedDirectionalMovementPlus[1]) - (nz(smoothedDirectionalMovementPlus[1])/adxLen) + directionalMovementPlus
    
    dIPlus := smoothedDirectionalMovementPlus / smoothedTrueRange * 100
    
    dIPlus
    
    
Adx(adxLen) =>
    dIPlus =  0.0
    dIMinus = 0.0
    dX = 0.0
    aDX = 0.0
    dIPlus := dIPlusCalc(adxLen)
    dIMinus := dIMinusCalc(adxLen)
    dX := abs(dIPlus-dIMinus) / (dIPlus+dIMinus)*100
    aDX := sma(dX, adxLen)
    
    aDX
    
BarInSession(sess) => time(timeframe.period, sess) != 0


//@version=4
strategy("Bollinger Band + RSI + ADX + MACD", overlay=true)

//Session

session = input(title="Trading Session", type=input.session, defval="0930-1500")

sessionColor = BarInSession(session) ? color.green : na

bgcolor(color=sessionColor, transp=95)

// Bollinger Bands
src = input(high, title="Bollinger Band Source", type=input.source)
length = input(3, minval=1, type=input.integer, title="Bollinger Band Length")
mult = input(4.989, minval=0.001, maxval=50, step=0.001, type=input.float, title="Bollinger Band Std Dev")
basis = sma(src, length)
dev = mult * stdev(src, length)
upper = basis + dev
lower = basis - dev


plot(upper, title="Bollinger Band Upper", color=color.red)
plot(lower, title="Bollinger Band Lower", color=color.green)

// RSI
rsiSrc = input(close, title="RSI Source", type=input.source)
rsiLength = input(16, minval=1, type=input.integer, title="RSI Length")
rsiComparator = input(39.2, title="RSI Comparator", type=input.float, step=0.1)

rsi = rsi(rsiSrc, rsiLength)

// ADX
adxLength = input(14, minval=1, type=input.integer, title="ADX Length")
adxComparator = input(14, minval=1, type=input.integer, title="ADX Comparator")

adx = Adx(adxLength)

// Heikinashi

haClose = security(heikinashi(syminfo.ticker), timeframe.period, close)
haOpen = security(heikinashi(syminfo.ticker), timeframe.period, open)

nextHaOpen = (haOpen + haClose) / 2

//MACD

macdCalcTypeProcessed = input(title="MACD Source", type=input.source, defval=high)
fast = input(12, title="MACD Fast")
slow = input(20, title="MACD Slow")
signalLen = input(15, title="MACD Signal")

fastMA = ema(macdCalcTypeProcessed, fast)
slowMA = ema(macdCalcTypeProcessed, slow)
macd = fastMA - slowMA
signal = sma(macd, signalLen)



longCondition() =>
    (low < lower) and (rsi[0] > rsiComparator) and (adx > adxComparator) and (close > nextHaOpen) and BarInSession(session) and macd > signal

stop = (close - max((low - (low * 0.0022)), (close - (close * 0.0032)))) / syminfo.mintick
target = (max(upper, (close + (close * 0.0075))) - close) / syminfo.mintick


strategy.entry("SX,LE", strategy.long, when=longCondition(), comment="SX,LE")
strategy.close_all(when=(not BarInSession(session)))
strategy.exit("LX", from_entry="SX,LE", profit=target, loss=stop)

Published by ChaoZhang on the FMZ strategy square · last modified 2024-02-01. 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/Quantitative-Trend-Tracking-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.