The resource loading... loading...

Price Divergence Based Trend Trading Strategy

Author: ChaoZhang, Date: 2024-02-02 18:00:55
Tags:

img

Overview

This is a trend trading strategy based on price divergence signals. It uses multiple indicators like RSI, MACD, Stochastics etc. to detect price divergences and the Murrey Math Oscillator to confirm. It enters when a price divergence signal appears and the oscillator confirms the current trend direction.

Strategy Logic

The core of this strategy is price divergence theory. When price reaches a new high but indicator doesn’t, it’s considered a bearish divergence. When price prints a new low but indicator doesn’t, it’s a bullish divergence. This signals a potential trend reversal. The strategy combines fractal signals with an oscillator to confirm trade signals.

Specifically, the entry conditions are:

  1. Detect regular/hidden price divergence
  2. Murrey Math Oscillator is in corresponding trend zone

Exit when the oscillator crosses middle line.

Advantage Analysis

The advantages of this strategy are:

  1. Detect potential reversal points using divergences
  2. Confirm ongoing trend with oscillator, avoiding false breakouts
  3. Flexible parameters and indicator combinations
  4. Combine trend following and risk management
  5. Clear logic rules, much room for optimization

Risk Analysis

The main risks are:

  1. Divergences could be false signals
  2. Improper oscillator parameters may cause missing trades
  3. Excessive one-sided positions bring large loss risk
  4. Increased trade frequency and slippage cost during high volatility periods

Suggest stop loss, position sizing, parameter optimization to reduce risks.

Optimization Directions

Some further optimizations:

  1. Add machine learning algorithms for dynamic parameter optimization
  2. Introduce more advanced stop loss techniques like trailing stop loss, average true range stop etc
  3. Incorporate more indicators and filters to improve signal-to-noise ratio
  4. Auto-adjust oscillator parameters for better trend judgement
  5. Enhance risk management, set maximum drawdown limits etc

Summary

This strategy integrates price divergence concept with trend analysis tools to discover potential reversals early. With proper risk management enhancements, it could achieve good risk-adjusted returns. Further machine learning based optimizations may lead to more stable alpha.


/*backtest
start: 2024-01-02 00:00:00
end: 2024-02-01 00:00:00
period: 2h
basePeriod: 15m
exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT"}]
*/

//@version=2
//
// Title:   [STRATEGY][UL]Price Divergence Strategy V1
// Author:  JustUncleL
// Date:    23-Oct-2016
// Version: v1.0
//
// Description:
//  A trend trading strategy the uses Price Divergence detection signals, that
//  are confirmed by the "Murrey's Math Oscillator" (Donchanin Channel based).
//
//  *** USE AT YOUR OWN RISK ***
//
// Mofidifications:
//  1.0 - original
//
// References:
//  Strategy Based on:
//  - [RS]Price Divergence Detector V2 by RicardoSantos
//  - UCS_Murrey's Math Oscillator by Ucsgears
//  Some Code borrowed from:
//  - "Strategy Code Example by JayRogers"  
//  Information on Divergence Trading:
//  - http://www.babypips.com/school/high-school/trading-divergences
//
strategy(title='[STRATEGY][UL]Price Divergence Strategy v1.0', pyramiding=0, overlay=true, initial_capital=10000, calc_on_every_tick=false,
         currency=currency.USD,default_qty_type=strategy.percent_of_equity,default_qty_value=10)
//  ||  General Input:
method = input(title='Method (0=rsi, 1=macd, 2=stoch, 3=volume, 4=acc/dist, 5=fisher, 6=cci):',  defval=1, minval=0, maxval=6)
SHOW_LABEL = input(title='Show Labels', type=bool, defval=true)
SHOW_CHANNEL = input(title='Show Channel', type=bool, defval=false)
uHid = input(true,title="Use Hidden Divergence in Strategy")
uReg = input(true,title="Use Regular Divergence in Strategy")
//  ||  RSI / STOCH / VOLUME / ACC/DIST Input:
rsi_smooth = input(title='RSI/STOCH/Volume/ACC-DIST/Fisher/cci Smooth:',  defval=5)
//  ||  MACD Input:
macd_src = input(title='MACD Source:', defval=close)
macd_fast = input(title='MACD Fast:',  defval=12)
macd_slow = input(title='MACD Slow:',  defval=26)
macd_smooth = input(title='MACD Smooth Signal:',  defval=9)
//  ||  Functions:
f_top_fractal(_src)=>_src[4] < _src[2] and _src[3] < _src[2] and _src[2] > _src[1] and _src[2] > _src[0]
f_bot_fractal(_src)=>_src[4] > _src[2] and _src[3] > _src[2] and _src[2] < _src[1] and _src[2] < _src[0]
f_fractalize(_src)=>f_top_fractal(_src) ? 1 : f_bot_fractal(_src) ? -1 : 0

//  ||••>   START MACD FUNCTION
f_macd(_src, _fast, _slow, _smooth)=>
    _fast_ma = sma(_src, _fast)
    _slow_ma = sma(_src, _slow)
    _macd = _fast_ma-_slow_ma
    _signal = ema(_macd, _smooth)
    _hist = _macd - _signal
//  ||<••   END MACD FUNCTION

//  ||••>   START ACC/DIST FUNCTION
f_accdist(_smooth)=>_return=sma(cum(close==high and close==low or high==low ? 0 : ((2*close-low-high)/(high-low))*volume), _smooth)
//  ||<••   END ACC/DIST FUNCTION

//  ||••>   START FISHER FUNCTION
f_fisher(_src, _window)=>
    _h = highest(_src, _window)
    _l = lowest(_src, _window)
    _value0 = .66 * ((_src - _l) / max(_h - _l, .001) - .5) + .67 * nz(_value0[1])
    _value1 = _value0 > .99 ? .999 : _value0 < -.99 ? -.999 : _value0
    _fisher = .5 * log((1 + _value1) / max(1 - _value1, .001)) + .5 * nz(_fisher[1])
//  ||<••   END FISHER FUNCTION

method_high = method == 0 ? rsi(high, rsi_smooth) : 
  method == 1 ? f_macd(macd_src, macd_fast, macd_slow, macd_smooth) :
  method == 2 ? stoch(close, high, low, rsi_smooth) :
  method == 3 ? sma(volume, rsi_smooth) :
  method == 4 ? f_accdist(rsi_smooth) :
  method == 5 ? f_fisher(high, rsi_smooth) :
  method == 6 ? cci(high, rsi_smooth) :
  na
    
method_low = method == 0 ? rsi(low, rsi_smooth) :
  method == 1 ? f_macd(macd_src, macd_fast, macd_slow, macd_smooth) :
  method == 2 ? stoch(close, high, low, rsi_smooth) :
  method == 3 ? sma(volume, rsi_smooth) :
  method == 4 ? f_accdist(rsi_smooth) :
  method == 5 ? f_fisher(low, rsi_smooth) :
  method == 6 ? cci(low, rsi_smooth) :
  na

fractal_top = f_fractalize(method_high) > 0 ? method_high[2] : na
fractal_bot = f_fractalize(method_low) < 0 ? method_low[2] : na

high_prev = valuewhen(fractal_top, method_high[2], 1) 
high_price = valuewhen(fractal_top, high[2], 1)
low_prev = valuewhen(fractal_bot, method_low[2], 1) 
low_price = valuewhen(fractal_bot, low[2], 1)

regular_bearish_div = fractal_top and high[2] > high_price and method_high[2] < high_prev
hidden_bearish_div = fractal_top and high[2] < high_price and method_high[2] > high_prev
regular_bullish_div = fractal_bot and low[2] < low_price and method_low[2] > low_prev
hidden_bullish_div = fractal_bot and low[2] > low_price and method_low[2] < low_prev

plot(title='H F', series=fractal_top ? high[2] : na, color=regular_bearish_div or hidden_bearish_div ? maroon : not SHOW_CHANNEL ? na : silver, offset=-2)
plot(title='L F', series=fractal_bot ? low[2] : na, color=regular_bullish_div or hidden_bullish_div ? green : not SHOW_CHANNEL ? na : silver, offset=-2)
plot(title='H D', series=fractal_top ? high[2] : na, style=circles, color=regular_bearish_div or hidden_bearish_div ? maroon : not SHOW_CHANNEL ? na : silver, linewidth=3, offset=-2)
plot(title='L D', series=fractal_bot ? low[2] : na, style=circles, color=regular_bullish_div or hidden_bullish_div ? green : not SHOW_CHANNEL ? na : silver, linewidth=3, offset=-2)

plotshape(title='+RBD', series=not SHOW_LABEL ? na : regular_bearish_div ? high[2] : na, text='R', style=shape.labeldown, location=location.absolute, color=maroon, textcolor=white, offset=-2)
plotshape(title='+HBD', series=not SHOW_LABEL ? na : hidden_bearish_div ? high[2] : na, text='H', style=shape.labeldown, location=location.absolute, color=maroon, textcolor=white, offset=-2)
plotshape(title='-RBD', series=not SHOW_LABEL ? na : regular_bullish_div ? low[2] : na, text='R', style=shape.labelup, location=location.absolute, color=green, textcolor=white, offset=-2)
plotshape(title='-HBD', series=not SHOW_LABEL ? na : hidden_bullish_div ? low[2] : na, text='H', style=shape.labelup, location=location.absolute, color=green, textcolor=white, offset=-2)

// Code borrowed from UCS_Murrey's Math Oscillator by Ucsgears
//  - UCS_MMLO
// Inputs
length = input(100, minval = 10, title = "MMLO Look back Length")
quad   = input(2, minval = 1, maxval = 4, step = 1, title = "Mininum Quadrant for MMLO Support")
mult = 0.125

// Donchanin Channel
hi = highest(high, length)
lo = lowest(low, length)
range = hi - lo
multiplier = (range) * mult
midline = lo + multiplier * 4

oscillator = (close - midline)/(range/2)

a = oscillator > 0
b = oscillator > 0 and oscillator > mult*2
c = oscillator > 0 and oscillator > mult*4
d = oscillator > 0 and oscillator > mult*6

z = oscillator < 0
y = oscillator < 0 and oscillator < -mult*2
x = oscillator < 0 and oscillator < -mult*4
w = oscillator < 0 and oscillator < -mult*6


//  Strategy: (Thanks to JayRogers)
// === STRATEGY RELATED INPUTS ===
//tradeInvert     = input(defval = false, title = "Invert Trade Direction?")
// the risk management inputs
inpTakeProfit   = input(defval = 0, title = "Take Profit Points", minval = 0)
inpStopLoss     = input(defval = 0, title = "Stop Loss Points", minval = 0)
inpTrailStop    = input(defval = 100, title = "Trailing Stop Loss Points", minval = 0)
inpTrailOffset  = input(defval = 0, title = "Trailing Stop Loss Offset Points", minval = 0)

// === RISK MANAGEMENT VALUE PREP ===
// if an input is less than 1, assuming not wanted so we assign 'na' value to disable it.
useTakeProfit   = inpTakeProfit  >= 1 ? inpTakeProfit  : na
useStopLoss     = inpStopLoss    >= 1 ? inpStopLoss    : na
useTrailStop    = inpTrailStop   >= 1 ? inpTrailStop   : na
useTrailOffset  = inpTrailOffset >= 1 ? inpTrailOffset : na

// === STRATEGY - LONG POSITION EXECUTION ===
enterLong() => ((uReg and regular_bullish_div) or (uHid and hidden_bullish_div)) and (quad==1? a[1]: quad==2?b[1]: quad==3?c[1]: quad==4?d[1]: false)// functions can be used to wrap up and work out complex conditions
exitLong() => oscillator <= 0
strategy.entry(id = "Buy", long = true, when = enterLong() )// use function or simple condition to decide when to get in
strategy.close(id = "Buy", when = exitLong() )// ...and when to get out

// === STRATEGY - SHORT POSITION EXECUTION ===
enterShort() => ((uReg and regular_bearish_div) or (uHid and hidden_bearish_div)) and (quad==1? z[1]: quad==2?y[1]: quad==3?x[1]: quad==4?w[1]: false)
exitShort() => oscillator >= 0
strategy.entry(id = "Sell", long = false, when = enterShort())
strategy.close(id = "Sell", when = exitShort() )

// === STRATEGY RISK MANAGEMENT EXECUTION ===
// finally, make use of all the earlier values we got prepped
strategy.exit("Exit Buy", from_entry = "Buy", profit = useTakeProfit, loss = useStopLoss, trail_points = useTrailStop, trail_offset = useTrailOffset)
strategy.exit("Exit Sell", from_entry = "Sell", profit = useTakeProfit, loss = useStopLoss, trail_points = useTrailStop, trail_offset = useTrailOffset)


//EOF

More