June 2009

TRADESTATION: Average True Range Trailing Stops

Sylvain Vervoort’s article in this issue, “Average True Range Trailing Stops,” describes a technique for generating trading signals with average true range calculations. Once the initial entry is made, any number of reversals may follow. The strategy’s first trade date and first trade direction are established by user inputs.

To download the EasyLanguage code for this study, go to the TradeStation and EasyLanguage Support Forum (https://www.tradestation.com/Discussions/forum.aspx?Forum_ID=213). Search for the file “Vervoort Atr Trail.eld.”

This article is for informational purposes. No type of trading or investment recommendation, advice, or strategy is being made, given or in any manner provided by TradeStation Securities or its affiliates.

Figure 1: TRADESTATION, ATR trailing stop strategy. Here is an example of the Vervoort ATR_Trail strategy on a chart of Google, Inc. (GOOG). The levels at which the strategy enters new long positions are displayed by the cyan lines. The short entry levels are displayed by the magenta lines. The chart on the left displays the levels based on the original ATR trail calculation. In the chart at right, the modified calculations are used.

Strategy: Vervoort ATR_Trail

{ Modified ATR Trailing Stop }
inputs:
	TrailType ( 1 ), { enter 1 for modified version, any
	 other number for unmodified version }
	ATR_Period( 5 ),
	ATR_Factor( 3.5 ),
	Quantity( 100 ),
	InitialMonth( 1 ),
	InitialDay( 1 ),
	InitialYear( 2009 ),
	FirstTrade( 1 ) ; { enter 1 for long, any other
	 number for short }

variables:
	Loss( 0 ),
	HiLo( 0 ),
	HRef( 0 ),
	LRef( 0 ),
	HiLoHRefMax( 0 ),
	HiLoHRefMaxLRefMax( 0 ),
	ATRMod( 0 ),
	WaitingForEntry( true ),
	Trail( 0 ),
	LineNum( 0 ),
	ReturnVal( 0 ) ;
 
if TrailType <> 1 then
	Loss = ATR_Factor * AvgTrueRange( ATR_Period )
else
	begin
	HiLo = iff( High - Low < 1.5 * Average( High - Low,
	 ATR_Period ), High - Low, 1.5 * Average( High -
	 Low, ATR_Period ) ) ;
	HRef = iff( Low <= High[1], High - Close[1],( High -
	 Close[1] ) - 0.5 * ( Low - High[1] ) ) ;
	LRef = iff( High >= Low[1], Close[1] - Low,
	 ( Close[1] - Low ) - 0.5 * ( Low[1] - High ) ) ;
	HiLoHRefMax = Maxlist( HiLo, HRef ) ;
	HiLoHRefMaxLRefMax = Maxlist( HiLoHRefMax, LRef ) ;
	ATRMod = XAverage( HiLoHRefMaxLRefMax, 2 *
	 ATR_Period - 1 ) ;
	Loss = ATR_Factor * ATRMod ;
	end ;

if WaitingForEntry 
	and Year( Date ) + 1900 >= InitialYear
	and Month( Date ) >= InitialMonth 
	and DayOfMonth( Date ) >= InitialDay 
then
	begin
	if FirstTrade = 1 then 
		begin
		Buy Quantity shares this bar Close ;
 	    WaitingForEntry = false ;
		Trail = Close - Loss ;
		end 
	else
		begin
		Sell short Quantity shares this bar at Close ; 
	 	WaitingForEntry = false ;
		Trail = Close + Loss ;
		end ;
	end 
else if WaitingForEntry[1] = false then
	begin
	if Close > Trail[1] and Close[1] > Trail[2] then
		{ continued long }		
		Trail = MaxList( Trail[1], Close - Loss )
	else if Close < Trail[1] and Close[1] < Trail[2]
	 then
		{ continued short }
		Trail = MinList( Trail[1], Close + Loss )
	else if Close > Trail[1] then  
		{ close is above trail }
		Trail = Close - Loss 
	else 
		Trail = Close + Loss ;

	if MarketPosition = -1 and Close > Trail and
	 Trail > 0 then
		begin
		Buy Quantity shares this bar Close ;
		LineNum = TL_New( Date[1], Time[1], Trail[1],
		 Date, Time, Trail[1] ) ;
		ReturnVal = TL_SetColor( LineNum, Cyan ) ;
		end 
	else if MarketPosition = 1 and Close < Trail then
		begin
		Sell short Quantity shares this bar at Close ;
		LineNum = TL_New( Date[1], Time[1], Trail[1],
		 Date, Time, Trail[1] ) ;
		ReturnVal = TL_SetColor( LineNum, Magenta ) ;
		end 
	else if Trail[1] > 0 then
		begin
		LineNum = TL_New( Date[1], Time[1], Trail[1],
		 Date, Time, Trail ) ;
		if Close > Trail then
			ReturnVal = TL_SetColor( LineNum, Magenta )
		else
			ReturnVal = TL_SetColor( LineNum, Cyan ) ;
		end ;
	end ;

—Mark Mills
TradeStation Securities, Inc.
A subsidiary of TradeStation Group, Inc.
www.TradeStation.com

BACK TO LIST

WEALTH-LAB: Average True Range Trailing Stops

The modified Atr indicator presented by Sylvain Vervoort in “Average True Range Trailing Stops” in this issue is available in Wealth-Lab’s TascIndicator library version 1.0.7.0 and later.

Here, we’ll provide the WealthScript code to enter a discretionary position, long or short, on the open of a specified date given by the strategy parameters. Included is the AtrTrail class, which utilizes the modified Atr to calculate a trailing stop based on the most-recent close. You can use it with any of your strategies. Once you create an AtrTrail object, just pass the bar number to the AtrTrail.Price method. As with last month’s article, if you prefer to use touch stops, WealthScript’s built-in ExitAtTrailingStop method is convenient.

See Figure 3 for a sample chart.

Figure 3: WEALTH-LAB, modified average true range stop. INTC’s three black crows following a doji-island reversal made a nice setup to enter short with a relatively tight stop at the center of the pattern.

WealthScript Code (C#): 

using System;
using System.Collections.Generic;
using System.Text;
using System.Drawing;
using WealthLab;
using WealthLab.Indicators;

namespace WealthLab.Strategies
{
   /* Class that encapsulates the ATR Trailing Stop, closing basis */
   public class ATRTrail
   {
      private WealthScript ws;
      private Bars bars;
      private double stopPrice;
      private bool longPosition;
      private DataSeries atrMod;
      
      public ATRTrail(WealthScript wL, Bars b, bool positionLong, double 
      initialStop, int period, double factor )
      {
         ws = wL;
         bars = b;
         longPosition = positionLong;
         stopPrice = initialStop;
         atrMod = factor * TASCIndicators.ATRModified.Series(bars, period);
      }
      
      // Call this method to update and return the stop price on each bar 
      after entry
      public double Price(int bar)
      {
         double prevPrice = stopPrice;
         double newPrice;         
         if (longPosition)
         {
            newPrice = bars.Close[bar] - atrMod[bar];
            stopPrice = newPrice > stopPrice ? newPrice : stopPrice;
         }
         else
         {
            newPrice = bars.Close[bar] + atrMod[bar];
            stopPrice = newPrice < stopPrice ? newPrice : stopPrice;
         }
         ws.DrawLine(ws.PricePane, bar-1, prevPrice, bar, stopPrice, Color.Blue,
          LineStyle.Solid, 1); 
         return stopPrice;
      }
   }
   
   public class SAC_ATRTrailingStops : WealthScript
   {
      private StrategyParameter _isLong = null;
      private StrategyParameter _initStop = null;
      private StrategyParameter _period = null;
      private StrategyParameter _atrMult = null; 
      private StrategyParameter _y = null;
      private StrategyParameter _m = null;
      private StrategyParameter _d = null;
      
      public SAC_ATRTrailingStops()
      {
         _isLong = CreateParameter("Long = 1", 1, 0, 1, 1); 
         _initStop = CreateParameter("Initial Stop", 1.0, 0.25, 50.0, 0.25); 
         _period = CreateParameter("ATR Period", 5, 2, 100, 1);
         _atrMult = CreateParameter("ATR Multiplier", 3.5, 1.0, 5.0, 0.1); 
         _m = CreateParameter("Month", 4, 1, 12, 1);
         _d = CreateParameter("Day", 13, 1, 31, 1);
         _y = CreateParameter("Year", 2009, 1990, 2012, 1); 
      }
      
      /* Execute a strategy - trade on a specified date */
      protected override void Execute()
      {      
         DateTime dt;         
         try {
            dt = new DateTime(_y.ValueInt, _m.ValueInt, _d.ValueInt); 
         }
         catch {
            DrawLabel(PricePane, "Invalid Date", Color.Red);
            return;
         }
         
         int b = Bars.ConvertDateToBar(dt, false);
         if (b < 1) {
            DrawLabel(PricePane, "Date does not exist on chart", Color.Red);
            return;
         }
         
         if(  _isLong.ValueInt == 1 )   
            BuyAtMarket(b, "Discretionary");
         else
            ShortAtMarket(b, "Discretionary");
         Position p = LastPosition;
         
         // After creating a position, initialize a stop object
         ATRTrail atrStop = new ATRTrail(this, Bars, p.PositionType ==
          PositionType.Long, _initStop.Value, _period.ValueInt, _atrMult.Value);
         
         for(int bar = b + 1; bar < Bars.Count; bar++)
         {            
            if (p.Active)
            {         
               if (p.PositionType == PositionType.Long)
               {
                  if( Close[bar] < atrStop.Price(bar) )
                     ExitAtMarket(bar + 1, p);
               }
               else if( Close[bar] > atrStop.Price(bar) )
                  ExitAtMarket(bar + 1, p); 
            } 
         } 
      }
   }
}

—Robert Sucher
www.wealth-lab.com

BACK TO LIST

NEUROSHELL TRADER: Average True Range Trailing Stops

The average true range trailing stop technique described by Sylvain Vervoort in his article in this issue, “Average True Range Trailing Stops,” can be implemented in NeuroShell Trader by combining a few of NeuroShell Trader’s built-in indicators plus the trading strategy wizard. First, to recreate J. Welles Wilder’s exponential average–based average true range indicator, select “New Indicator…” from the Insert menu and use the Indicator Wizard to create the following indicator:

Average true range (ATR) indicator:
ExpAvg( Subtract ( Max2(High, Lag(Close,1)), Min2( Low, Lag(Close,1))), 9 )

To recreate Sylvain Vervoort’s average true range trailing stop reversal system, select “New Trading Strategy…” from the Insert menu and enter the following in the appropriate locations of the Trading Strategy Wizard:

Long protective stop:
Subtract( MaxValueSinceEntryFilled(TradingStrategy, High, 1), 
Multiply2( 3.5, WilderAverageTrueRange))
 
Short protective stop:
Add2( MinValueSinceEntryFilled(TradingStrategy, Low, 1), 
Multiply2( 3.5, WilderAverageTrueRange))

If you wish to create an average true range trailing stop system using your own trading system’s entry rules, simply combine the trailing stops listed above with your own entry/exit conditions.

To recreate the modified average true range indicator, select “New Indicator…“ from the Insert menu and use the Indicator Wizard to create the following indicators:

HiLo 
Min2(Subtract(High,Low), MovAvg(Subtract(High,Low),5))
 
Href
Subtract(Subtract(High, Lag(Close,1)), IfThenElse( A<=B(Low, Lag(High,1)), 
0, Divide(Subtract(Low, Lag(High,1)), 2)))
 
Lref
Subtract(Subtract(Lag(Close,1), Low), IfThenElse ( A>=B(High, Lag(Low,1)), 
0, Divide(Subtract(Lag(Low,1), High), 2)))
 
Modified Average True Range
ExpAvg( Max3( HiLo, Href, Lref ), 9)

To recreate the modified average true range trailing stop reversal system, enter the following formula in the appropriate locations of the Trading Strategy Wizard:

Long protective stop:
Subtract( MaxValueSinceEntryFilled(TradingStrategy #2, High, 1), 
Multiply2( 3.5, ModifiedAverageTrueRange))
 
Short protective stop:
Add2( MinValueSinceEntryFilled(TradingStrategy #2, Low, 1), 
Multiply2( 3.5, ModifiedAverageTrueRange))

If you wish to create a modified average true range trailing stop system using your own trading system’s entry rules, simply combine the modified average true range trailing stops listed above with your own entry/exit conditions.

Figure 5: NeuroShell TRADER, AVERAGE TRUE RANGE TRAILING STOP SYSTEM

If you have NeuroShell Trader Professional, you can also choose whether the system parameters should be optimized. After backtesting the trading strategies, use the “Detailed Analysis…” button to view the backtest and trade-by-trade statistics for each strategy.

For more information on NeuroShell Trader, visit www.NeuroShell.com.

—Marge Sherald, Ward Systems Group, Inc.
301 662-7950, sales@wardsystems.com
www.neuroshell.com

BACK TO LIST

TradersStudio: Average True Range Trailing Stops

The TradersStudio implementation of the fixed-percentage trailing stop system and the date-specific version based on “Average True Range Trailing Stops” by Sylvain Vervoort is discussed here.

The average true range (Atr) trailing stop should have an advantage over the fixed-percentage trailing stop that was tested last month in my May 2009 Traders’ Tip, since the stop can adapt to the changing volatility of both the overall market as well as to each individual stock’s volatility and changes in volatility.

In my May 2009 Traders’ Tip, I modified Vervoort’s fixed-percentage trailing stop system by adding market timing and also by trading both the long and short sides. This system, if traded both long and short, would be always in, but since I added a market-timing filter based on a general-market trend-following filter, it will only trade long when the market trend is up or short when the market trend is down.

The coded version of Vervoort’s Atr trailing stop system that I am supplying here has the options of trading either long only, short only, or both long and short. It also has the option to apply a market-trend filter using the S&P 500 index (Spx) to determine whether to trade long or short. I decided to stick with the author’s list of stocks for the tests.

One advantage in testing stocks with TradersStudio over other programs is that both the unadjusted price series as well as the split-adjusted series are available in the tests. This can make a big difference in the computation of commissions when they are based on the number of shares traded. (Other programs compute commissions on a cents-per-share basis, and unless you know the actual price of the stock and the actual number of shares that would have been traded, commissions cannot be computed accurately.) In addition, with TradersStudio, we are able to correctly apply a minimum price rule to eliminate very low-priced stocks. When only split-adjusted data is available, we cannot accurately apply a price filter because we don’t know whether a stock is low-priced due to a split or was really trading at a low price in real time. TradersStudio also has the ability to compute the dividends that you would have been received or paid (if short).

Figure 6: TITLE.blurb

In Figure 6, I show a comparison of the equity curves: the left one represents trading both long and short using a trend filter on the Spx with optimized parameters using the fixed-percentage trailing stop system from the May 2009 issue, while the one on the right shows this issue’s modified Atr trailing stop (AtrM) system with optimized parameters. Just by looking at the two equity curves, it’s hard to tell which is preferable, although the AtrM (right curve) appears smoother with smaller drawdowns. Looking at the table in Figure 7, which shows some key metrics for the two trailing-stop methods, we see that the AtrM stops are preferable, since we obtain slightly more net profit with lower drawdown, better profit factor, and better profit-to-maximum drawdown ratios with fewer trades. I also found that there is a slight improvement (not shown) by using the AtrM system versus the Atr system.

Figure 7: TITLE.blurb

To measure the robustness of the parameter sets from the AtrM system optimization, in Figure 8 I show two three-dimensional models. The left model shows the two parameters compared to the net profit, and the right model shows the two parameters compared to the maximum drawdown. The Atr-length parameter is less sensitive to changes than the Atr multiple. The range of good parameters is five to 10 days for the Atr length, and 4.5 to 5.5 for the Atr multiple. All tests used 300 days for the moving average length on the Spx trend filter. I used a TradersStudio add-in to produce the three-dimensional models.

Figure 8: TITLE.blurb

The code can be downloaded from the TradersStudio website at www.TradersStudio.com ->Traders Resources->FreeCode as well as from www.TradersEdgeSystems.com/traderstips.htm. It can also be copied and pasted from the Stocks & Commodities website at www.traders.com.

' AVERAGE TRUE RANGE TRAILING STOPS (ATR and ATRM)
' Author: Sylvain Vervoort, TASC June 2009
' Coded by: Richard Denning 4/8/09
' www.tradersEdgeSystems.com

Sub ATR_TSSR(tsStartDate,WilderLenATR,multATR,longOnly,useMktTm,spxLen,useModATR)
'tsStartDate = date that trading is to start after bars back has been satisfied
'           use TradeStation date format of YYYMMDD example 1/20/2003 = 1030120
'WilderLenATR = 5, multATR = 3.5, unitSiz = 1, longOnly = 2 (both long and short)
'longOnly = 1 (trade long side only), longOnly = -1 (short side only)
'useMktTm = 1 will apply an SPX trend filter so that trades are taken only in
'           the direction of the trend of the SP500 index; any other value
'           and no market timing filter is applied
' spxLen = 250
' useModATR = 1 uses modified ATR; any other value uses standard ATR
Dim unitSiz
Dim trail As BarArray
Dim rsAIQst As BarArray
Dim size As BarArray
Dim SPXc As BarArray
Dim OKtoBuy, OKtoSell
unitSiz = 1
SPXc = C Of independent1
OKtoBuy = SPXc > Average(SPXc,spxLen,0) And SPXc[1] > Average(SPXc,spxLen,1)
OKtoSell = SPXc < Average(SPXc,spxLen,0) And SPXc[1] < Average(SPXc,spxLen,1)
If useModATR = 1 Then
    trail = ATRM_TRAIL_STOP(WilderLenATR, multATR)
Else
    trail = ATR_TRAIL_STOP(WilderLenATR, multATR)
End If

size = (1000 / TSCLose) / unitSiz

If Not CrossesOver(C,trail,0) And Not CrossesUnder(C,trail,0) Then size = size[1]
If useMktTm = 1 Then
    If Date >= MigrateDate(tsStartDate) Then
        If longOnly >= 1 And OKtoBuy Then
            If CrossesOver(C,trail,0) Then Buy("LE",size,0,CloseEntry,Day)
            If CrossesUnder(C,trail,0) Then ExitLong("LX","",size[1],0,
            CloseExit,Day)
        End If 
        If (longOnly <= 0 Or longOnly = 2) And OKtoSell Then
            If CrossesUnder(C,trail,0) Then Sell("SE",size,0,CloseEntry,Day)
            If CrossesOver(C,trail,0) Then ExitShort("SX","",size[1],0,
            CloseExit,Day)
        End If
    End If
Else
    If Date >= MigrateDate(tsStartDate) Then
        If longOnly >= 1 Then
            If CrossesOver(C,trail,0) Then Buy("LE",size,0,CloseEntry,Day)
            If CrossesUnder(C,trail,0) Then ExitLong("LX","",size[1],0,
            CloseExit,Day)
        End If 
        If (longOnly <= 0 Or longOnly = 2) Then
            If CrossesUnder(C,trail,0) Then Sell("SE",size,0,CloseEntry,Day)
            If CrossesOver(C,trail,0) Then ExitShort("SX","",size[1],0,
            CloseExit,Day)
        End If
    End If
End If
End Sub

'_____________________________________________

Function ATR_Wilder(WilderLen)
' computes average true range using Wells Wilder averaging method
    ATR_Wilder = XAverage(TrueRange,WilderLen*2-1,0)
End Function

'_____________________________________________

Function ATR_TRAIL_STOP(WilderLen,multATR)
' suggested default parameters: WilderLen = 5, multATR = 3.5
Dim loss
Dim longStop
Dim shortStop
Dim maxVal
Dim minVal
Dim ATRval As BarArray
Dim trail As BarArray
loss = ATR_Wilder(WilderLen)*multATR
longStop = C - loss
shortStop = C + loss
maxVal = Max(trail[1],longStop)
minVal = Min(trail[1],shortStop)
If C>trail[1] And C[1]>trail[2] Then
    trail = maxVal
Else If Ctrail[1] And C[1]trail[2] Then
    trail = shortStop
Else
    trail = C
End If
End If
End If
End If
    
ATR_TRAIL_STOP = trail

End Function

'_____________________________________________

Function ATRM_Wilder(WilderLen)
' computes average true range as modified using Wells Wilder averaging method
    Dim AvgRng As BarArray
    Dim HiLo As BarArray
    Dim Href As BarArray
    Dim Lref As BarArray
    Dim HLmax As BarArray
    Dim LHmax As BarArray
    Dim Diff1 As BarArray
    Dim Diff2 As BarArray
    
    AvgRng = Average(H - L, WilderLen)    
    HiLo = IIF(H - L < 1.5 * AvgRng, H - L, 1.5*AvgRng)
    HLmax = (H - C[1]) - (L - H[1]) / 2
    Href = IIF(L <= H[1],H - C[1],HLmax)
    LHmax = (C[1] - L) - (L[1] - H) / 2
    Lref = IIF(H >= L[1], C[1] - L, LHmax)
    Diff1 = Max(HiLo,Href)
    Diff2 = Max(Diff1,Lref)
    ATRM_Wilder = XAverage(Diff2,WilderLen*2-1,0)
End Function

'_____________________________________________

Function ATRM_TRAIL_STOP(WilderLen,multATR)
' coputes the modified ATR trailing stop
' suggested default parameters: WilderLen = 5, multATR = 3.5
Dim loss
Dim longStop
Dim shortStop
Dim maxVal
Dim minVal
Dim ATRval As BarArray
Dim trail As BarArray
loss = ATRM_Wilder(WilderLen)*multATR
longStop = C - loss
shortStop = C + loss
maxVal = Max(trail[1],longStop)
minVal = Min(trail[1],shortStop)
If C>trail[1] And C[1]>trail[2] Then
    trail = maxVal
Else If Ctrail[1] And C[1]trail[2] Then
    trail = shortStop
Else
    trail = C
End If
End If
End If
End If
  
ATRM_TRAIL_STOP = trail

End Function

'_____________________________________________

'ATR Trailing Stop Indicator
sub ATR_TRAIL_STOP_IND(WilderLen,multATR)
    plot1(ATR_TRAIL_STOP(WilderLen,multATR))
End Sub

'_____________________________________________

'ATR Modified Trailing Stop Indicator
Sub ATRM_TRAIL_STOP_IND(WilderLen,multATR)
    plot1(ATRM_TRAIL_STOP(WilderLen,multATR))
End Sub

'_____________________________________________

—Richard Denning
for TradersStudio
richard.denning@earthlink.net

BACK TO LIST

STOCKFINDER: Average True Range Trailing Stops

We’ve created a chart for StockFinder that shows plots for the fixed-percentage stop, average true range trailing stop (Atr TS), modified Atr TS, and a modified Atr, based on ’Average True Range Trailing Stops“ by Sylvain Vervoort. The standard average true range is already available in the indicator library.

To load the chart into any layout, click the Shared Items icon (envelope) at the top of the StockFinder screen then choose “Browse other users’ shared items” from the menu that appears. Click the Charts tab from the Shared Items window that appears and search for “June Traders’ Tips.” Open the chart by selecting it from the list and clicking the Open button. The chart will open in your current layout. You can save the chart by clicking the Save button at the top of the chart.

The red plot on price is the fixed-percentage stop from Vervoort’s article. Clicking it brings up its Edit window. From there, you can change the month, day, and year for the trade open, the stop, and the percentage. The blue plot on price is the Atr TS from the article. Clicking it brings up its Edit window. From there you can change the month, day, and year for the trade open, the stop, Atr period, and multiple. The yellow plot on price is the modified Atr TS from the article. Clicking it brings up its Edit window. From there you can change the month, day, and year for the trade open, the stop, Atr period, and multiple. The bottom pane shows the modified Atr plot, which can also be edited by clicking on it.

Rules to scan, sort, or color plots can be created for any plot by clicking on it and using the appropriate tab from the Edit window. If you have any questions about the chart or plots, visit the Worden support forums at www.Worden.com/Training.

For more information on the software or to start your free trial, visit www.StockFinder.com.

Figure 10: STOCKFINDER, ATR TRAILING STOP. This chart demonstrates the fixed-percentage stop, average true range trailing stop (ATR TS), modified ATR TS, and a modified ATR.

—Craig Shipman
Worden Brothers, Inc
www.Worden.com, www.StockFinder.com

BACK TO LIST

TRADECISION: Average True Range Trailing Stops

In his series of articles on stop methods, Sylvain Vervoort demonstrates the importance of considering a warning signal and setting stops at the right time to avoid getting stopped out by the normal noise of the market. This issue’s article, “Average True Range Trailing Stops,” describes a trailing-stop method based on the average true range (Atr) trailing stop.

Using the Function Builder in Tradecision, create the Stop_Trail_ATR_Mod function for a modified Atr trailing stop value as follows:

function (Period:numeric=5,atrfact:Numeric=3.5):Numeric;
var
 HiLo:=0;
 Href:=0;
 Lref:=0;
 diff1:=0;
 diff2:=0;
 atrmod:=0;
 loss:=0;
 result:=0;
end_var

   HiLo:=iff(H - L < 1.5 * Mov(H - L, period, S), H - L, 1.5 * Mov(H - L,
     period, S));
   Href:=iff(L <= Ref(H, -1), H - Ref(C, -1), (h - ref(C, -1)) - (L -
     Ref(H, -1)) / 2);
   Lref:=iff(H >= Ref(L, -1), Ref(C, -1) - L, (Ref(C, -1) - L) - (Ref(L, -1)
    - H) / 2);

   diff1:=Max(HiLo, Href);
   diff2:=Max(diff1, Lref);

   atrmod:=EMA(diff2,period);
   loss:=atrfact*atrmod;

if HISTORYSIZE <= period then result := C;
else
  begin
    if C > this\1\ AND C\1\ > this\1\  then result := Max(this\1\,C-loss);
    else
    begin
       if C < this\1\ then result := Min(this\1\,C+loss);
       else
       begin
          if C > this\1\ then result := C-loss;
          else result := C+loss;
       end;
    end;
  end;
return result;
 
Then you specify the strategy rules in Tradecision's Strategy Builder:
Entry Long:
if Date() = 080102 then return true;
return false;

Exit Long:
return CrossBelow(Stop_Trail_ATR_Mod(5,3.5),C);
Entry Short:
if Date() = 081201 then return true;
return false;

Exit Short:
return CrossAbove(Stop_Trail_ATR_Mod(5,3.5),C);

Figure 12: TRADECISION, MODIFIED AVERAGE TRUE RANGE TRAILING STOP. Here is a 3.5 times five-period modified ATR average plotted on a chart of DD.

As mentioned in Vervoort’s article, you’ll need to enter the starting date manually. To define Date(), use a numeric value in the Yymmdd format. Date returns “080102” if the day is January 2, 2008. See Figure 12.

To import the strategy into Tradecision, visit the area “Traders’ Tips from Tasc magazine” at https://tradecision.com/support/tasc_tips/tasc_traders_tips.htm or copy the code from the Stocks & Commodities website at www.traders.com.

—Yana Timofeeva
Alyuda Research, Inc.
510 931-7808, sales@tradecision.com
www.tradecision.com

BACK TO LIST

WAVE59: Average True Range Trailing Stops

In “Average True Range Trailing Stops” in this issue, Sylvain Vervoort implements an Atr stop algorithm that he uses as both a reversal system and as a way to apply a trailing stop to open positions entered using other methods.

We were interested to see how Vervoort’s trailing stop method handled recent action in the volatile stock indexes, so we applied it to the Dow Jones Industrial Average. You can see in Figure 14 that it stayed short for the latter half of 2008 and reversed to a long position in March 2009. According to this tool, the market looks ready to regain some of the losses posted last year. Just be wary if we cross back underneath the red line!

Figure 14: WAVE59, ATR TRAILING STOP. Here is Vervoort’s trailing stop method applied to the Dow Jones Industrial Average. It stayed short for the latter half of 2008 and reversed to a long position in March 2009.

We are offering four scripts that implement Vervoort’s approach in Wave59. As always, users of Wave59 can download these scripts directly using the QScript Library, found at https://www.wave59.com/library.

Indicator 1: ATR reversal system
#This script plots Sylvain Vervoort's ATR Stop Value
#as described in the June 2009 issue of Stocks&Commodities
#Period: period to get ATR value, using wilder's average function
#Multiplier: factor to multiply the ATR value by to calculate stop amount
#Act_as_System: if true, use as trading system, otherwise just lines
input:period(5),multiplier(3.5),act_as_system(false),color(red),thickness(1);

#initialize variables
if (barnum==barsback) {
   stopval=close;
   buysell=1;
}

#calculate atr values
atr=wilder_average(truerange(),period);
bigatr=atr*multiplier;

#calculate long stops
if (buysell>0) {
   stopval=close-bigatr;
   stopval=max(stopval,stopval[1]);
   if (close<stopval) {
      buysell=-1;
      stopval=close+bigatr;
      if (act_as_system==true) exitlong(0,close,"market","onebar");
   }
}

#calculate stort stops
else if (buysell<0) {
     stopval=close+bigatr;
     stopval=min(stopval,stopval[1]);
     if (close>stopval) {
        buysell=1;
        stopval=close-bigatr;
        if (act_as_system==true) buy(1,close,"market","onebar");
     }
}

#plot the lines
plot1=stopval;
color1=color;
thickness1=thickness;

 
Indicator 2: ATR system using a fixed date

#This script plots Sylvain Vervoort's ATR Stop Value
#as described in the June 2009 issue of Stocks&Commodities.
#Use this version to manage an actual trade based on a specific date.
#Period: period to get ATR value, using wilder's average function
#Multiplier: factor to multiply the ATR value by to calculate stop amount
#Startyear, Startmonth, Startday: Day to Begin plotting on
input:startyear(2009),startmonth(1),startday(1),longposition(true),
   period(5),multiplier(3.5),color(red),thickness(1);

#initialize variables
if (barnum==barsback) {
   startbar=date_to_bar(startyear,startmonth,startday,time);
   stopval=close;
   buysell=1;
   if (longposition==false) buysell=-1;
   txref=text(barnumber, close, "", color, tx_left, 8);
}

#calculate atr values
atr=wilder_average(truerange(),period);
bigatr=atr*multiplier;

if (barnum<startbar)
{
   if (buysell>0) stopval=close-bigatr;
   else stopval=close+bigatr;
}
else
{
    #calculate long stops
    if (buysell>0) {
       stopval=close-bigatr;
       stopval=max(stopval,stopval[1]);
       if (close<stopval) {
          buysell=-1;
          stopval=close+bigatr;
       }
    }

    #calculate stort stops
    else if (buysell<0) {
         stopval=close+bigatr;
         stopval=min(stopval,stopval[1]);
         if (close>stopval) {
            buysell=1;
            stopval=close-bigatr;
         }
    }

    #plot the lines
    plot1=stopval;
    color1=color;
    thickness1=thickness;

    #plot a label
    text_setbar(txref,barnum+5);
    text_setprice(txref,stopval);
    text_setstring(txref,to_string(stopval));
}
 
Indicator 3: This script uses Vervoort's Modified ATR calculation

#This script plots Sylvain Vervoort's Modified ATR Stop Value
#as described in the June 2009 issue of Stocks&Commodities
#Period: period to get modified ATR value
#Multiplier: factor to multiply the ATR value by to calculate stop amount
#Act_as_System: if true, use as trading system, otherwise just lines
input:period(5),multiplier(3.5),act_as_system(false),color(red),thickness(1);

#initialize variables
if (barnum==barsback) {
   stopval=close;
   buysell=1;
}

#calculate Vervoort's modified atr values
simplehilo=average(high-low,period);
if (high-low<1.5*simplehilo)
    hilo=high-low;
else
    hilo=simplehilo;
href=abs(high-close[1]);
if (low>high[1]) href-=(low-close[1])/2;
lref=abs(low-close[1]);
if (high<low[1]) lref-=(close[1]-high)/2;
modatr=max(lref,href,hilo);
bigatr=wilder_average(modatr,period);
bigatr*=multiplier;

#calculate long stops
if (buysell>0) {
   stopval=close-bigatr;
   stopval=max(stopval,stopval[1]);
   if (close<stopval) {
      buysell=-1;
      stopval=close+bigatr;
      if (act_as_system==true) exitlong(0,close,"market","onebar");
   }
}

#calculate stort stops
else if (buysell<0) {
     stopval=close+bigatr;
     stopval=min(stopval,stopval[1]);
     if (close>stopval) {
        buysell=1;
        stopval=close-bigatr;
        if (act_as_system==true) buy(1,close,"market","onebar");
     }
}

#plot the lines
plot1=stopval;
color1=color;
thickness1=thickness;
 
Indicator 4: This script uses Vervoort's Modified ATR calculation to
 manage a specific trade

#This script plots Sylvain Vervoort's ATR Stop Value
#as described in the June 2009 issue of Stocks&Commodities.
#Use this version to manage an actual trade based on a specific date.
#Period: period to get ATR value, using wilder's average function
#Multiplier: factor to multiply the ATR value by to calculate stop amount
#Startyear, Startmonth, Startday: Day to Begin plotting on
input:startyear(2009),startmonth(1),startday(1),longposition(true),
   period(5),multiplier(3.5),color(red),thickness(1);

#initialize variables
if (barnum==barsback) {
   startbar=date_to_bar(startyear,startmonth,startday,time);
   stopval=close;
   buysell=1;
   if (longposition==false) buysell=-1;
   txref=text(barnumber, close, "", color, tx_left, 8);
}

#calculate Vervoort's modified atr values
simplehilo=average(high-low,period);
if (high-low<1.5*simplehilo)
    hilo=high-low;
else
    hilo=simplehilo;
href=abs(high-close[1]);
if (low>high[1]) href-=(low-close[1])/2;
lref=abs(low-close[1]);
if (high<low[1]) lref-=(close[1]-high)/2;
modatr=max(lref,href,hilo);
bigatr=wilder_average(modatr,period);
bigatr*=multiplier;

if (barnum<startbar)
{
   if (buysell>0) stopval=close-bigatr;
   else stopval=close+bigatr;
}
else
{
    #calculate long stops
    if (buysell>0) {
       stopval=close-bigatr;
       stopval=max(stopval,stopval[1]);
       if (close<stopval) {
          buysell=-1;
          stopval=close+bigatr;
       }
    }

    #calculate stort stops
    else if (buysell<0) {
         stopval=close+bigatr;
         stopval=min(stopval,stopval[1]);
         if (close>stopval) {
            buysell=1;
            stopval=close-bigatr;
         }
    }

    #plot the lines
    plot1=stopval;
    color1=color;
    thickness1=thickness;

    #plot a label
    text_setbar(txref,barnum+5);
    text_setprice(txref,stopval);
    text_setstring(txref,to_string(stopval));
}


--------------------------------------------

—Earik Beann
Wave59 Technologies Int’l, Inc.
www.wave59.com

BACK TO LIST

TRADE-IDEAS: Average True Range Trailing Stops

“Ideas are easy; it’s execution that’s hard.”—Jeff Bezos (Amazon.com)

In “Average True Range Trailing Stops” in this issue, Sylvain Vervoort creates a trailing stop based on the average true range value on a portfolio of 25 stocks and pursues the question of whether it’s a better stop than the fixed-percentage trailing stop method demonstrated in his May 2009 article. Subscribers to Trade-Ideas Pro who use our event-based backtesting tool, The OddsMaker, already understand the different premise these tools use to find high-probability winning trades. To recap briefly, The OddsMaker doesn’t just look at 25 stocks, à priori, to generate backtest results, it considers any stock that matched a desired pattern in the market, finds that stock, and applies the backtest’s rule set before summing up the results into a detailed set of totals: win rate, average winner, average loser, net winnings, and confidence factor. (See Figure 18.) As such, we provide an alternative stop method and another strategy for your consideration.

The wiggle
At Trade-Ideas we use a custom, dynamic stop that takes into account a stock’s 15-minute average volatility multiplied by the relative volume (an indexed ratio of what the current volume is over its historical volume of the stock at the time a trade is made). This measure is called the wiggle.

The wiggle can be further explained with an example. Nflx’s 15-minute average volatility is $0.3306/15 minutes. If the relative volume of Nflx at the time of an alert is 1.5 (that is, trading at 50% above what it normally trades at this time of the day of the alert based on accumulated volume of the day), then the wiggle is $0.50 (that is, $0.3306*1.5). Look up the volatility values of any stock at the Trade-Ideas.com Stock Research section. Imagine using a $0.50 stop and attempting to stay in a GS short or long; chances are, you will be correct about the trade but get stopped out, only to watch the stock do exactly what you thought it would, which we all know is painful. The wiggle creates a customized stop that takes into account the characteristics of each stock.

The wiggle is used in the strategy, “Modified Atr Volatility Stop” and is based on the Trade-Ideas inventory of alerts and filters found in our flagship product, Trade-Ideas Pro. The trading rules are modeled and backtested in its add-on tool, The OddsMaker.

Description: “Modified ATR Volatility Stop” 
Provided by:
Trade Ideas (copyright © Trade Ideas LLC 2009). All rights reserved. 
For educational purposes only. Remember these are sketches meant to 
give an idea how to model a trading plan. Trade-Ideas.com and all 
individuals affiliated with this site assume no responsibilities 
for trading and investment results.

Figure 16: TRADE-IDEAS, ALERTS CONFIGURATION. The combination of alerts and filters used to create Sylvain Vervoort’s modified ATR volatility stop strategy are shown here.

Type or copy/paste this shortened string directly into a browser then copy/paste the full length link into Trade-Ideas PRO using the “Collaborate” feature (right-click in any strategy window):

https://bit.ly/1WrpZ (case sensitive)

This strategy also appears on the Trade-Ideas blog, https://marketmovers.blogspot.com/, or you can build the strategy from Figure 16, which shows the configuration of this strategy, where one alert and nine filters are used with the following settings:

• Running Up Now (in 1 minute or less), $0.75
• Min Price Filter = 5 ($)
• Max Price Filter = 50 ($)

The definitions of these indicators appear here: https://www.trade-ideas.com/Help.html.

That’s the strategy, but what about the trading rules? How should the opportunities that the strategy finds be traded? For the stop-loss, we used one of the more advanced options available in the OddsMaker. Instead of using a traditional stop-loss, we selected “Another alert” as the exit condition called “Trailing stop, volatility down.” This alert is triggered when a stock moves down an amount equal to the average 15-minute volatility multiplied by the number of 15-minute bars we decide on — in this case, 2. For example, if the average 15-minute volatility is 10 cents, this alert would not trigger until a stock moves down at least 20 cents.

Here is what The OddsMaker tested for the past three weeks ended 4/9/2009 given the following trade rules:

• On each alert, sell short the symbol (price moves down to be a successful trade)
• Schedule an exit for the stocks 30 minutes after entry
• Start trading from the open and stop trading 30 minutes later
• Set a stop using the alert, trailing stop, volatility down, with a setting of two bars

The OddsMaker summary provides evidence of how well this strategy and our trading rules did. The settings are shown in Figure 17.

Figure 17: TRADE-IDEAS, ODDSMAKER. Here is the backtesting configuration for the modified ATR volatility stop strategy.

The results (last backtested for the three-week period ended 4/9/2009) are shown in Figure 18. The summary reads as follows: This strategy generated 498 trades of which 339 were profitable for a win rate of 68.07%. The average winning trade generated $0.38 in profit and the average loser lost $0.18. The net winnings of using this strategy for 15 trading days generated $105.06 points. If you normally trade in 100-share lots, this strategy would have generated $10,506. The z-score or confidence factor that the next set of results will fall within this strategy’s average winner and loser is 100%.

Figure 18: TRADE-IDEAS, RESULTS. Here are sample results from The Oddsmaker for the modified ATR volatility stop strategy.

You can find an explanation of The OddsMaker backtest results in more detail from the online user manual at https://www.trade-ideas.com/OddsMaker/Help.html.

—Dan Mirkin & David Aferiat
Trade Ideas, LLC
david@trade-ideas.com, www.trade-ideas.com

BACK TO LIST