Dynamic Breakout EURUSD by George Pruitt

• 10/22/2017 at 7:44 PM #50250
  Despair

  I stumbled over an article about a channel breakout system that varies the length of the channel depending on volatility.

  Here is the article:

  https://www.quantconnect.com/tutorials/dynamic-breakout-ii-strategy/

  I thought it was an interesting idea and programmed it. It produces quite good backtest for EURUSD on daily bars.

  There are not many trades but maybe this can be developed in a useful strategy. Varying the deviation of the bollinger bands would be definitely worth trying (i did not try that). I just took the rules from the articles and gave it a run – no optimization at all.

  I almost forgot, I made 1 change. The article uses the standard deviation of the lookback as measure for volatility. I had this first and this worked but then i tried ATR instead as measure of volatility and results improved. The code to use standard deviation instead of ATR is there still as comment.

  Dynamic Breakout EURUSD

  defparam cumulateorders=false once lookback=20 once positionsize=1 atr=AverageTrueRange[lookback](close) tradingtime=(currentdayofweek>=1 and currentdayofweek<=5) If tradingtime then todayvol=atr //STD[lookback](close) ydayvol=atr[1] //STD[lookback](close[1]) deltavol=(todayvol-ydayvol)/todayvol lookback=round(lookback*(1+deltavol)) lookback=MIN(60,lookback) lookback=MAX(20,lookback) MA=Average[lookback](close) BoUp=BollingerUp[lookback](close) BoDown=BollingerDown[lookback](close) LC1=close[1]>BoUp LC2=close>highest[lookback](high[1]) SC1=close[1]<BoDown SC2=close<lowest[lookback](low[1]) if LC1 and LC2 then buy positionsize contract at market elsif SC1 and SC2 then sellshort positionsize contract at market endif if longonmarket and (close < MA ) then sell at market elsif shortonmarket and (close > MA ) then exitshort at market endif endif

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

  defparam cumulateorders=false   once lookback=20 once positionsize=1   atr=AverageTrueRange[lookback](close)   tradingtime=(currentdayofweek>=1 and currentdayofweek<=5)   If tradingtime then   todayvol=atr //STD[lookback](close)   ydayvol=atr[1] //STD[lookback](close[1])   deltavol=(todayvol-ydayvol)/todayvol   lookback=round(lookback*(1+deltavol))   lookback=MIN(60,lookback)   lookback=MAX(20,lookback) MA=Average[lookback](close) BoUp=BollingerUp[lookback](close) BoDown=BollingerDown[lookback](close) LC1=close[1]>BoUp LC2=close>highest[lookback](high[1]) SC1=close[1]<BoDown SC2=close<lowest[lookback](low[1]) if LC1 and LC2 then    buy positionsize contract at market elsif SC1 and SC2 then    sellshort positionsize contract at market endif if longonmarket and (close < MA ) then    sell at market elsif shortonmarket and (close > MA ) then    exitshort at market endif endif

   

  3 users thanked author for this post.

  Nicolas, victormork, Berra

  

  Despair

  Blocked

  Topics: 97

  Replies: 1034

  Been thanked: 217 times
  10/29/2017 at 10:34 AM #50815
  Nicolas

  Thank you Despair for sharing this code with us. Could be a nice addition to the library too, just for learning purpose and maybe helping others ideas come to us. Should work with other forex pairs with some WFO for validation?

  Apparently, the good trade (big breakout) are greatly helping the equity curve to climb up! False signals came from ranging market and it can still long times like months/years 🙁

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20828

  Been thanked: 4742 times
  10/29/2017 at 5:55 PM #50829
  Despair

  Definitely. This is nothing you want to trade as is. But I thought the idea is interesting and maybe somebody wants to proceed to work with it. I did not put much work/time in it.

  

  Despair

  Blocked

  Topics: 97

  Replies: 1034

  Been thanked: 217 times
• Author
  Posts