Profitable strategy that work on any market

• 08/02/2017 at 5:00 PM #42343
  Henrik

  @juanj, no I forgot to du that. But now i tried to change but I dont get the same results. Here is 200k from 8 to 19 utc+1.

  

  Henrik

  Participant

  Topics: 9

  Replies: 97

  Been thanked: 19 times
  08/02/2017 at 9:23 PM #42371
  juanj

  @Henrik and @Nicolas I have spent some time ‘optimizing’ Entry and Exit criteria, possibly to the extent where it is ‘over’ optimized (too many OR’s). I will over the comming days refine the criteria adding and removing them as I test between different markets trying to find what tend to work more universally. For now the attached is the latest version, also attached my latest results on EUR/USD. @Hendrik if you backtest on 200k bars just re-optimize the trading time filter using the backtest engine.

  

  juanj

  Participant

  Topics: 51

  Replies: 505

  Been thanked: 170 times
  08/03/2017 at 10:01 AM #42388
  Henrik

  Here it is over 200k. Put it optimizing  overnight. Feels like stopped somewhere but you see which time parameters used.

  A=starttime

  S=endtime

  D=exit

   

  Regards

   

  Henrik

  1 user thanked author for this post.

  Nicolas

  

  Henrik

  Participant

  Topics: 9

  Replies: 97

  Been thanked: 19 times
  08/03/2017 at 10:08 AM #42390
  juanj

  @Hendrik, thankyou for that I am glad to see it is performing well going back that far. Looking really positive!

  I would say according to your optimization Start: 9 and End: 19 is the sweet spot here, good 10 hours of exposure and the highest win %.

  The Exit time is not really too important here as it is simply meant to close the positions Friday night (just before the weekend).

  

  juanj

  Participant

  Topics: 51

  Replies: 505

  Been thanked: 170 times
  08/03/2017 at 2:19 PM #42406
  juanj

  I have developed a new ‘universal’ strategy developed on many of the same principles as the other one using a Keltner band to supplement the Bollinger band and trading mean reversion.

  It is not meant to replace the previous strategy but rather to supplement it.

  I have also not tested this strategy on as many markets as my other one however it performs relatively fair across markets as far as I could see so far. Attached is the .ITF file along with the results on Gold (attached screenshot compares the 2 strategies result on Gold). @Henrik if you want you can maybe test the strategy on 200k bars optimizing the timeframe accordingly (just Start Time and End Time)?

  
  @Nicolas if you prefer we can add this strategy as a separate entry to the library?

   

  

  juanj

  Participant

  Topics: 51

  Replies: 505

  Been thanked: 170 times
  08/03/2017 at 2:58 PM #42411
  Nicolas

  @juanj

  I think we could make multiple posts but with different approach, because it seems to evolve each day a bit more 🙂 People may also “fork” this code and upload them here. So I feel like we could wait for things to be a bit more ‘advanced’ or ‘completed’. In any case, thanks for the job done and for sharing! 😉 Please note that I’ll be on leave by tomorrow, not be able to be as present on forums/admin and moderations as usual.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  08/03/2017 at 3:08 PM #42414
  juanj

  @nicolas enjoy your leave! When will you be back in full swing?

  

  juanj

  Participant

  Topics: 51

  Replies: 505

  Been thanked: 170 times
  08/03/2017 at 6:43 PM #42440
  Henrik

  Hi!

  Here is BollingerKeltner.itf with 200k and time opti.

  Regards

  Henrik

  

  Henrik

  Participant

  Topics: 9

  Replies: 97

  Been thanked: 19 times
  08/03/2017 at 6:51 PM #42442
  Inertia

  Bonnes vacances Nicolas !!

  

  Inertia

  Participant

  Topics: 40

  Replies: 223

  Been thanked: 52 times
  08/04/2017 at 8:00 AM #42463
  juanj

  @Hendrik thank you for that. What do you say we try a few more markets? Oil as well as some of the other big markets? FTSE, DAX, etc.

  

  juanj

  Participant

  Topics: 51

  Replies: 505

  Been thanked: 170 times
  08/06/2017 at 7:35 PM #42570
  juanj

  I have modified my latest version to include heuristics to automatically optimize the period used to calculate the Bollinger and Keltner Bands. I have also opened a new topic (containing the heuristics algorithm used).

  https://www.prorealcode.com/topic/machine-learning-in-proorder/

  

  juanj

  Participant

  Topics: 51

  Replies: 505

  Been thanked: 170 times
  08/07/2017 at 8:35 AM #42629
  Marcel van Vliet

  The sequence make no sense to me. I compared it with the mathematical sequences I could find, did some calculations in Excel and also visit a mathematical forum for more info. The only thing I could confirm is the number 42 from @juanj that seems to be the median of the sequence.

  

  Marcel van Vliet

  Participant

  Topics: 62

  Replies: 98

  Been thanked: 7 times
  08/07/2017 at 9:34 AM #42636
  juanj

  @Marcel thank you for your comment, I appreciate you looking into it. You are correct in that the sequence doesn’t make sense, however it does achieve its purpose. Maybe you can assist with a more logical algorithm? What I am trying to achieve is something like this;

  Starting Variable (Median)

  Starting Variable + (increment)

  Starting Variable –  (increment)

  Starting Variable + (increment*2)

  Starting Variable – (increment*2)

  Starting Variable + (increment*3)

  Starting Variable – (increment*3

  Etc.

  However in typing this I now think I know how to do it.

  

  juanj

  Participant

  Topics: 51

  Replies: 505

  Been thanked: 170 times
  08/07/2017 at 10:15 AM #42640
  juanj

  //Heuristics Algorithm Increment = 1 once IncPos = 1 //Increment Position once Optimize = 0 once Mode = 1 //Switches between + and - increments once StratAvgB = 0 once Periods = 42 //Starting eriod of variable to optimize If Optimize = 10 Then //'optimize = optimize + 1' set at position entry/exit WinLoss = 0 StratAvgA = StratAvgB StratAvgB = 0 For i = 1 to 10 Do If positionperf(i) > 0 Then WinLoss = WinLoss + 1 EndIF StratAvgB = StratAvgB + StrategyProfit[i] Next StratAvgB = StratAvgB/10 If (Winloss < 5 or StratAvgA > StratAvgB) and Mode = 1 Then Periods = Periods + (Increment*IncPos) ElsIf (Winloss < 5 or StratAvgA > StratAvgB) and Mode = 2 Then Periods = Periods - (Increment*IncPos) mode = 1 If IncPos > 6 Then //Max Increment Value Periods = 42 //Starting Vaue is Reset IncPos = 1 //Increment Position is Reset Else IncPos = IncPos + 1 EndIF EndIF mode = 2 Optimize = 0 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 35

  //Heuristics Algorithm   Increment = 1 once IncPos = 1 //Increment Position once Optimize = 0 once Mode = 1 //Switches between + and - increments once StratAvgB = 0 once Periods = 42 //Starting eriod of variable to optimize   If Optimize = 10 Then //'optimize = optimize + 1' set at position entry/exit WinLoss = 0 StratAvgA = StratAvgB StratAvgB = 0 For i = 1 to 10 Do If positionperf(i) > 0 Then WinLoss  = WinLoss + 1 EndIF StratAvgB = StratAvgB + StrategyProfit[i] Next StratAvgB = StratAvgB/10 If (Winloss < 5 or StratAvgA > StratAvgB) and Mode = 1 Then Periods = Periods + (Increment*IncPos) ElsIf (Winloss < 5 or StratAvgA > StratAvgB) and Mode = 2 Then Periods = Periods - (Increment*IncPos) mode = 1 If IncPos > 6 Then //Max Increment Value Periods = 42 //Starting Vaue is Reset IncPos = 1 //Increment Position is Reset Else IncPos = IncPos + 1 EndIF EndIF mode = 2 Optimize = 0 EndIf

   

  

  juanj

  Participant

  Topics: 51

  Replies: 505

  Been thanked: 170 times
  08/07/2017 at 10:36 AM #42651
  volpiemanuele

  @ Juani

  I’m still working to the strategy based on Ichimoku and hosoda cycle. If you are interested we can develop togheter a specific strategy.

  Thanks

  Emanuele

  

  volpiemanuele

  Participant

  Topics: 35

  Replies: 189

  Been thanked: 49 times
• Author
  Posts