Discussing the strategy VECTORIAL DAX (M5)

• 05/10/2020 at 4:12 PM #130956
  Paul

  Here is my modification of Balmora’s original. I went back to the beginning and made a few small changes with in mind it had to be a lite version.

  Besides that, the trailing stop kicks in when atr distance is 8 and quickly moves to 4 when there are no new high/low made in an effort not to surrender too much gains.

  With 2% stoploss and a trailingstop there are lot’s off signals which aren’t visible when in a position. So it’s important to be able to include those, especially in the optimisation!

  Inserted reenter, so every true conditions to enter the market could be tested, even when current position was in loss, profit or both. (selectable) Then it’s closed and opened again in same direction.

  It means more trades and smaller average losses/gains and still this strategy is maintaining a similar result despite more trades.  a good indication for robustness!

   

  10 users thanked author for this post.

  Dr Manhattan, Edmond, GraHal, Balmora74, keewee, Sebastian Arsjo, OboeOpt, nonetheless, Francesco, Nacho

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/10/2020 at 9:34 PM #131031
  Exalaxe

  Little question which should be obvious for you:

  How many occurencies should i choose in the Walk Forward?

  

  Exalaxe

  Participant

  Topics: 0

  Replies: 4

  Been thanked: 0 times
  05/10/2020 at 10:52 PM #131041
  swedshare

  If it’s the 121-version you wish to backtest you can set it to 1 and then inactive.

  

  swedshare

  Participant

  Topics: 2

  Replies: 66

  Been thanked: 30 times
  05/11/2020 at 8:45 PM #131203
  Exalaxe

  ok. Thank you.

  

  Exalaxe

  Participant

  Topics: 0

  Replies: 4

  Been thanked: 0 times
  05/12/2020 at 1:12 PM #131300
  Franro

  todays position wasnt good at all,,

  franro.

  

  Franro

  Participant

  Topics: 25

  Replies: 219

  Been thanked: 23 times
  05/13/2020 at 12:54 PM #131493
  Nacho

  Good premiere on the first day it operates!

  

  Nacho

  Participant

  Topics: 0

  Replies: 23

  Been thanked: 10 times
  05/13/2020 at 5:44 PM #131535
  Franro

  Nacho you jinxed it

  

  Franro

  Participant

  Topics: 25

  Replies: 219

  Been thanked: 23 times
  05/15/2020 at 2:26 PM #131778
  Paul

  one thing to point out.

  running it for long only and short only, both performances with current settings are poor. Long & short work together with eachother and both have in part optimised settings. That is one major reliability issue regardless how the performance looks if both are active at the same time, is it not?

   

   

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/16/2020 at 7:24 PM #131885
  Paul

  on 121p, simply selecting long/short only resulted in (very) poor performance.

  It uses the stoploss of 2% and a big trailingstop.

  So new test, the focus is long/short only, it needed a smaller stoploss and changed it to 0.5%

  Made the trailingstop faster (atr 6), optimised the settings (and went back to the original values) and optimised the angle. Without the trend detection.

   

   

   

   

  5 users thanked author for this post.

  GraHal, Balmora74, Sebastian Arsjo, swedshare, nonetheless

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/17/2020 at 12:53 PM #131954
  Balmora74

  Hi Paul. What do you mean when you say : “without the trend detection” ? . It is the Average [20] and/ or Average[35] condition for CONDBUY and CONDSELL ?

  

  Balmora74

  Participant

  Topics: 5

  Replies: 97

  Been thanked: 78 times
  05/17/2020 at 1:51 PM #131965
  Paul

  Hi, yeah that’s it.

  here’s the latest, I don’t know how to improve with only few extra criteria significantly for only short & only long.

   

   

   

   

   

  7 users thanked author for this post.

  von, nonetheless, Balmora74, Scooby, GraHal, swedshare, keewee

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/17/2020 at 10:41 PM #132037
  Balmora74

  Paul this is the last version of the code i’m running since one month and the results seems quite good. I added ATR + Daily / Weekly and Monthly Pivot for calculating the PositionSize and as a filter for entry point. Perhaps a idea to improve the your excellent v130p version of Vectorial Dax…

   

   

   

  8 users thanked author for this post.

  von, vcharlesc, BobFlynn, swedshare, GraHal, zilliq, nonetheless, Paul

  

  Balmora74

  Participant

  Topics: 5

  Replies: 97

  Been thanked: 78 times
  05/18/2020 at 1:22 AM #132050
  Paul

  Thanks a lot for releasing a new version Balmora74! Looks very interesting, first thing I think where v130 can be improved with is pivots. Got some work to do! 🙂

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/20/2020 at 11:47 PM #132490
  BobFlynn

  Maybe it’s curve fitting guys, but I get best results when using 5 instead of 11 and 20 instead of 36 for the ATR position sizing. Tested the robustness and it looks quite alright, let me know what you think 🙂

  
  @Balmora74, how did you found the ATR values ? I’m working on a more gradual sizing position on your “basic” version.

  Anyhow, even older versions are working quite well despite the crash, I’m very surprised, i honestly thought there was some overfitting in the air, but apparently not so much ! Good job guys!

  1 user thanked author for this post.

  swedshare

  

  BobFlynn

  Participant

  Topics: 15

  Replies: 44

  Been thanked: 10 times
  05/21/2020 at 10:55 AM #132557
  zilliq

  Paul wrote:

  Hi, yeah that’s it.

  here’s the latest, I don’t know how to improve with only few extra criteria significantly for only short & only long.

   

  Hi @Paul
  Why do you separate long/short robots ? Do you think it ameliorate something ?

  Cheers

   

  

  zilliq

  Participant

  Topics: 58

  Replies: 683

  Been thanked: 152 times
• Author
  Posts