Discussing the strategy VECTORIAL DAX (M5)

• 09/01/2019 at 12:28 PM #106069
  Nacho

  Good morning, in case I miss something that I have not taken into account, why don’t you use version 4.2? For me it is the one that gives the best results with a very tight DD. At the end of July I put in real and I am getting very good results, attached screenshot.

  

  Nacho

  Participant

  Topics: 0

  Replies: 23

  Been thanked: 10 times
  09/02/2019 at 10:14 AM #106200
  Stefanb

  According to my back test, V5 is better with less DD.

  In addition, it has performed really well since it was made in April

  

  Stefanb

  Participant

  Topics: 15

  Replies: 74

  Been thanked: 22 times
  09/02/2019 at 10:17 AM #106202
  Stefanb

  I use size 0.5.

  Even with 1 contract, V5 has lower DD..

  But does V2 make a bigger profit?

  

  Stefanb

  Participant

  Topics: 15

  Replies: 74

  Been thanked: 22 times
  09/02/2019 at 6:43 PM #106265
  Conny Gustafsson

  @Stefanb what settings are you running? usetimecriteria etc?

  

  Conny Gustafsson

  Participant

  Topics: 0

  Replies: 2

  Been thanked: 0 times
  09/03/2019 at 5:50 PM #106384
  nfulcher

  I’ve just finished reading this thread – looks like a great strategy.

  I’m still finding my way around PRT and IG and would like to confirm something. Regarding the backtesting done by various people above, is that on the Germany 30 ‘CFD’ account or ‘Spreadbet’ account (and is there a way to tell from the posted results)?

  Thanks.

   

   

  

  nfulcher

  Participant

  Topics: 1

  Replies: 17

  Been thanked: 0 times
  09/03/2019 at 5:53 PM #106385
  Balmora74

  @nfulcher

  CFD Germany30 with IG Markets.

   

  1 user thanked author for this post.

  nfulcher

  

  Balmora74

  Participant

  Topics: 5

  Replies: 97

  Been thanked: 77 times
  09/03/2019 at 6:29 PM #106394
  Nacho

  Good, when I wrote I could not study version 5 quietly, but this is better, having a drowdown less than laitad that 4.2 is preferable to activate version 5.0 with 2 contracts to activate 4.2, since it would take less money and earn plus. Thanks for putting backtest with 200,000 candles.

  

  Nacho

  Participant

  Topics: 0

  Replies: 23

  Been thanked: 10 times
  09/03/2019 at 6:46 PM #106397
  Balmora74

  @Nacho

  Backtest with 200k candles / PositionSize = 1 / Spread = 1 with these variables :

  once tradeschedule = 0 //[0]=full; [1]=morning; [2]=lunch; [3]=afternoon
  once resetcounter = 1 //[0]=default; [1]=experiment (both connected by position criteria)once enablesl = 1 // stoploss
  once enablept = 0 // profit target
  once enablets = 1 // trailing stop
  once enablebb = 0 // bollingerband exit (add bb as indicator to have it visually) (all can be combined)

  1 user thanked author for this post.

  BobFlynn

  

  Balmora74

  Participant

  Topics: 5

  Replies: 97

  Been thanked: 77 times
  09/03/2019 at 6:54 PM #106403
  AE

  Hi @Balmora74

  Thanks you for this fantastic strategy. Two questions:

  1. Which variables do you recomend to check in order to fit to the market (maybe the angle? the MAs?)
  2. How many ofen do you recomend check this variables?

  Thanks again!

  

  AE

  Participant

  Topics: 11

  Replies: 53

  Been thanked: 12 times
  09/03/2019 at 7:09 PM #106405
  Balmora74

  Thanks @AE

  QUESTION 1

  You can play with :

  ONCE PeriodeA = 10
  ONCE nbChandelierA= 15
  ONCE PeriodeB = 20
  ONCE nbChandelierB= 35
  ONCE lag = 1.5

  and also the variable “pente”

  In order to help you in this work you can had these 2 indicators

  B.M.I. (ANGLE) - Balmora74

  PeriodeA = 10 //Période de la MM nbChandelierA= 20 // Nombre de chandeliers sur lesquels on évalue la pente MM = Exponentialaverage[PeriodeA](close) //pente = (MM-MM[nbchandelierA])/nbchandelierA //trigger = Exponentialaverage[Periode](pente) ADJASUROPPO = (MM-MM[nbchandelierA]*pipsize) / nbChandelierA ANGLE = (ATAN(ADJASUROPPO)) RETURN angle as "Angle"

  1 2 3 4 5 6 7 8 9 10

  PeriodeA = 10  //Période de la MM nbChandelierA= 20 // Nombre de chandeliers sur lesquels on évalue la pente MM = Exponentialaverage[PeriodeA](close) //pente = (MM-MM[nbchandelierA])/nbchandelierA //trigger = Exponentialaverage[Periode](pente) ADJASUROPPO = (MM-MM[nbchandelierA]*pipsize) / nbChandelierA ANGLE = (ATAN(ADJASUROPPO))     RETURN angle as "Angle"

  B.M.I. (PENTE) - Balmora

  PeriodeB = 30 //Période de la MM nbChandelierB = 40 // Nombre de chandeliers sur lesquels on évalue la pente lag = -10 MM = Exponentialaverage[PeriodeB](close) pente = (MM-MM[nbchandelierB]*pointsize)/nbchandelierB trigger = Exponentialaverage[Periodeb + lag](pente) //ADJASUROPPO = (ABS(MM-MM[nbchandelierB]*pointsize)) / nbChandelierB //ANGLE = (ATAN(ADJASUROPPO)) RETURN Pente as "Pente", trigger as "Trigger"

  1 2 3 4 5 6 7 8 9 10 11 12

  PeriodeB = 30  //Période de la MM nbChandelierB = 40 // Nombre de chandeliers sur lesquels on évalue la pente lag = -10   MM = Exponentialaverage[PeriodeB](close) pente = (MM-MM[nbchandelierB]*pointsize)/nbchandelierB trigger = Exponentialaverage[Periodeb + lag](pente) //ADJASUROPPO = (ABS(MM-MM[nbchandelierB]*pointsize)) / nbChandelierB //ANGLE = (ATAN(ADJASUROPPO))     RETURN Pente as "Pente", trigger as "Trigger"

  See attached piece for the graphical configuration of these indicators.

  For your question number 2 = It’s a good question 🙂

   

   

   

   

   

  1 user thanked author for this post.

  Paul

  

  Balmora74

  Participant

  Topics: 5

  Replies: 97

  Been thanked: 77 times
  09/03/2019 at 9:30 PM #106409
  nfulcher

  Balmora74 wrote:

  @nacho

  Backtest with 200k candles / PositionSize = 1 / Spread = 1 with these variables :

  once tradeschedule = 0 //[0]=full; [1]=morning; [2]=lunch; [3]=afternoon

  once resetcounter = 1 //[0]=default; [1]=experiment (both connected by position criteria)once enablesl = 1 // stoploss

  once enablept = 0 // profit target

  once enablets = 1 // trailing stop

  once enablebb = 0 // bollingerband exit (add bb as indicator to have it visually) (all can be combined)

  I’ve backtested this tonight using 100,000 and got the following results. This is nowhere near the 95% gain you got above. Is this due to the backtest period.

  I’ve used the following variables:

  //strategy once periodea = 10 once nbchandeliera= 15 once periodeb = 25 //20 once nbchandelierb= 35 once lag = 1.5

  1 2 3 4 5 6

  //strategy once periodea = 10 once nbchandeliera= 15 once periodeb = 25 //20 once nbchandelierb= 35 once lag = 1.5

   

  

  nfulcher

  Participant

  Topics: 1

  Replies: 17

  Been thanked: 0 times
  09/04/2019 at 9:02 AM #106442
  tba69

  Google traduction :

  Hello Balmora74,

  thank you again for your strategy and the idea of ​​giving the corresponding indicators helps to better understand it.

  I test this strategy in real and demo, actually I also find some trade differences if the trailing stop is affected at the same time or not.

  Moreover, I was able to test strategies in M1 and the problem can be very important with SL affected in real while demo we did a TP.

  In demo there are also overnight and overweek charges that are not taken into account.

  Regarding the stop follower, I noticed that often the gain realized is well below the maximum gain possible on a position (in the table MFE).

  Is there the possibility to modify the variables of this trailing stop which is based on the ATR?

  Have a good day.

  

  tba69

  Participant

  Topics: 1

  Replies: 4

  Been thanked: 0 times
  09/04/2019 at 9:28 AM #106448
  Nicolas

  @tba69

  Only english speaking in english forums sections of the website please.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20828

  Been thanked: 4738 times
  09/04/2019 at 12:06 PM #106473
  Stefanb

  I have had different versions of this code live but now I am considering increasing size of contracts.

  Have some € 100 in profit on this code and might as well bet the “profit”

  Now I use v5.

  Will start with 3 contracts.

  Do you have the code live or on demo?

  

  Stefanb

  Participant

  Topics: 15

  Replies: 74

  Been thanked: 22 times
  09/04/2019 at 1:04 PM #106477
  GraHal

  @Stefanb … might be better to wait until you are £1000 in profit and then bet £100 of the free money?

  That way you will still be happy even if it goes wrong?? 🙂

  I am still in Demo with Vectorials, but I will be  going Live as soon as PRT v11 is provided by IG.

  If anybody wants to enter performance in the Log below for any version – Forward Test on Demo or Live using live data – then we may all help each other?

  Forward Test – Sys Perf

   

   

  1 user thanked author for this post.

  Balmora74

  

  GraHal

  Participant

  Topics: 203

  Replies: 8390

  Been thanked: 2035 times
• Author
  Posts