Mean Reversion Trading System (EUR/USD 1H)

• 11/21/2017 at 6:52 PM #53490
  Gustaf

  Hi everyone!

  I’m quite new to this, but have been lurking here for a while.
  This is a mean reversion trading system made for EUR/USD on the hourly chart that goes both long and short.

  If the high touches the upper bollinger band and the stochastic crosses under a fixed value, the system goes short (vice versa for the long side).
  Exits are made when stochastics crosser over a fixed value.

  I used an ATR for filtering out bad trades.
  I found it more efficient to use a trendfilter for the long positions (high > close[300]). For the shorts, no trendfilter is used (this concerns me).

  The system seems to perfom good on 100k units – but since I am a rookie I can’t imagine that this will get any good results on a 200k backtest.
  Would anyone mind testing it on 200k and post results?

  I’m all ears if you have any suggestions for improvement.

  Bollinger - Mean Reversion - EUR/USD 1H

  // DEFINITION OF CODE PARAMETERS DEFPARAM CumulateOrders = True DEFPARAM PRELOADBARS = 500 // CONDITIONS TO ENTER SHORT indicator1 = Stochastic[14,3](close) c1 = (indicator1 CROSSES UNDER 85) indicator2 = BollingerUp[50](high) c2 = (high >= indicator2) indicator3 = AverageTrueRange[14](close) c3 = indicator3 >= (0.00125) IF c1 AND c2 AND c3 AND NOT SHORTONMARKET THEN SELLSHORT 1 CONTRACT AT MARKET ENDIF // CONDITIONS TO EXIT SHORT indicator4 = Stochastic[14,3](close) c4 = (indicator4 CROSSES OVER 10) IF c4 THEN EXITSHORT AT MARKET ENDIF // CONDITIONS TO ENTER LONG PosTrend = high > close[300] indicator5 = Stochastic[14,3](close) c5 = (indicator5 CROSSES OVER 10) indicator6 = BollingerDown[50](close) c6 = (low <= indicator6) indicator7 = AverageTrueRange[14](close) c7 = indicator7 >= (0.00125) IF PosTrend AND c5 AND c6 AND c7 AND NOT LONGONMARKET THEN BUY 1 CONTRACT AT MARKET ENDIF // CONDITIONS TO EXIT LONG indicator8 = Stochastic[14,3](close) c8 = (indicator8 CROSSES OVER 75) IF c8 THEN SELL AT MARKET ENDIF //STOP/TARGET SET STOP PLOSS 150

  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 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53

  // DEFINITION OF CODE PARAMETERS   DEFPARAM CumulateOrders = True DEFPARAM PRELOADBARS = 500     // CONDITIONS TO ENTER SHORT   indicator1 = Stochastic[14,3](close) c1 = (indicator1 CROSSES UNDER 85) indicator2 = BollingerUp[50](high) c2 = (high >= indicator2) indicator3 = AverageTrueRange[14](close) c3 = indicator3 >= (0.00125)     IF c1 AND c2 AND c3 AND NOT SHORTONMARKET THEN SELLSHORT 1 CONTRACT AT MARKET ENDIF   // CONDITIONS TO EXIT SHORT indicator4 = Stochastic[14,3](close) c4 = (indicator4 CROSSES OVER 10)   IF c4 THEN EXITSHORT AT MARKET ENDIF   // CONDITIONS TO ENTER LONG PosTrend = high > close[300] indicator5 = Stochastic[14,3](close) c5 = (indicator5 CROSSES OVER 10) indicator6 = BollingerDown[50](close) c6 = (low <= indicator6) indicator7 = AverageTrueRange[14](close) c7 = indicator7 >= (0.00125)     IF PosTrend AND c5 AND c6 AND c7 AND NOT LONGONMARKET THEN BUY 1 CONTRACT AT MARKET ENDIF   // CONDITIONS TO EXIT LONG indicator8 = Stochastic[14,3](close) c8 = (indicator8 CROSSES OVER 75)   IF c8 THEN SELL AT MARKET ENDIF     //STOP/TARGET SET STOP PLOSS 150

  Prints of results below, tested with spread set to 0.8.

  

  Gustaf

  Participant

  Topics: 5

  Replies: 11

  Been thanked: 0 times
  11/21/2017 at 7:16 PM #53495
  AutoStrategist

  Not as good over 200k.  Strategies using Bolinger bands for mean reversion seem to work quite well on a one hour timeframe, I have one of my own.

  1 user thanked author for this post.

  Gustaf

  

  AutoStrategist

  Participant

  Topics: 6

  Replies: 125

  Been thanked: 71 times
  11/21/2017 at 7:21 PM #53499
  AutoStrategist

  Also just noticed I didn’t put any spread in.

  

  AutoStrategist

  Participant

  Topics: 6

  Replies: 125

  Been thanked: 71 times
  11/21/2017 at 7:23 PM #53500
  Gustaf

  Thank you for the backtest, really appreciate it!

   

  

  Gustaf

  Participant

  Topics: 5

  Replies: 11

  Been thanked: 0 times
  11/22/2017 at 9:03 AM #53545
  jebus89

  I mean, it doesnt look as bad if u just start backtest on 200K a little later hehehe.

   

  Im more sceptical of how few trades there are.. 200K backtest and u got like 304 trades.. and this long and short? Seems a bit too curve-fitted. Also numbers are just too high and pretty to be true 😀

  

  jebus89

  Participant

  Topics: 118

  Replies: 586

  Been thanked: 238 times
• Author
  Posts