timebased series Dax

• 10/25/2019 at 4:58 AM #111141
  JohnScher

  Based on a different behavior of buyers and sellers on monday morning compared to friday evening as a basic approach in a “timebased series”, there is a long positioning on tuesday evening in the Dax.

  Below you will find the code for a simple trading system that follows this approach. Similarly, this can already be found in the library as an advertisement for the vtad.

  Here, however, it was stated that “buy tuesday evening, sell thursday morning” can be a positive approach, especially if it is underlaid with only 2 simple filters, 1x an EMA filter, 1x a CCI filter.

  The result seems impressive.

  I wanted to ask if someone could check the code for correctness and if so if someone could check the strategy for a longer distance than the offered 100.000 units.

   

  buy tuesday evening, sell thursday morning

  //------------------------------------------------------------------------- // Main code : D2 18 1H Long H33 // created and coded by JohnScher //------------------------------------------------------------------------- defparam cumulateorders = true TradingDay = opendayofweek = 2 TradingTime = time = 180000 c1 = close < ExponentialAverage [30] (close) c2 = cci [24] (close) > -140 Tradebuy = TradingDay and TradingTime and c1 and c2 IF Tradebuy THEN buy at market ENDIF IF onmarket and barindex-tradeindex = 33 then // = D4 040000 notice the asian session sell at market Endif Set Stop %Loss 5 Set Target %Profit 2.5

  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

  //------------------------------------------------------------------------- // Main code : D2 18 1H Long H33 // created and coded by JohnScher //-------------------------------------------------------------------------   defparam cumulateorders = true     TradingDay  = opendayofweek = 2 TradingTime = time = 180000     c1 = close < ExponentialAverage [30] (close) c2 = cci [24] (close) > -140     Tradebuy = TradingDay and TradingTime and c1 and c2   IF Tradebuy  THEN buy at market ENDIF   IF onmarket and barindex-tradeindex = 33 then // = D4 040000 notice the asian session sell at market Endif   Set Stop %Loss 5 Set Target %Profit 2.5

   

   

  

  JohnScher

  Participant

  Topics: 45

  Replies: 338

  Been thanked: 87 times
  10/25/2019 at 9:59 AM #111161
  Nicolas

  Thanks John for sharing the idea.

  Here is the backtest on 200.000 units with 2 points spread. The robustness test on the 100k out of sample period is a proof of over fitting. You might have made too much optimization of your indicators settings?

  1 user thanked author for this post.

  JohnScher

  

  Nicolas

  Keymaster

  Topics: 83

  Replies: 20899

  Been thanked: 4868 times
  10/25/2019 at 12:22 PM #111184
  GraHal

  Hope you din’t mind John, I couldn’t help wondering what results would look if not Sell after 33 hours (top curve).  Food for thought maybe??

  What does the version below look like on 200k bars anybody please?

  //https://www.prorealcode.com/topic/timebased-series-dax/ //------------------------------------------------------------------------- // Main code : D2 18 1H Long H33 // created and coded by JohnScher //------------------------------------------------------------------------- defparam cumulateorders = False TradingDay = opendayofweek = 2 TradingTime = time = 170000 // c1 = close < ExponentialAverage [30] (close) c2 = cci [24] (close) > -150 //-140 Tradebuy = c1 and c2 and TradingDay and TradingTime IF Tradebuy THEN buy at market SET STOP PLOSS 350 SET TARGET PPROFIT 425 ENDIF //IF onmarket and barindex-tradeindex = 33 then // = D4 040000 notice the asian session //sell at market //Endif //Set Stop %Loss 5 //Set Target %Profit 2.5

  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

  //https://www.prorealcode.com/topic/timebased-series-dax/ //------------------------------------------------------------------------- // Main code : D2 18 1H Long H33 // created and coded by JohnScher //------------------------------------------------------------------------- defparam cumulateorders = False TradingDay  = opendayofweek = 2 TradingTime = time = 170000 // c1 = close < ExponentialAverage [30] (close) c2 = cci [24] (close) > -150 //-140 Tradebuy =  c1 and c2 and TradingDay and TradingTime IF Tradebuy  THEN buy at market SET STOP PLOSS 350 SET TARGET PPROFIT 425 ENDIF //IF onmarket and barindex-tradeindex = 33 then // = D4 040000 notice the asian session //sell at market //Endif //Set Stop %Loss 5 //Set Target %Profit 2.5

   

  1 user thanked author for this post.

  JohnScher

  

  GraHal

  Participant

  Topics: 222

  Replies: 9296

  Been thanked: 2462 times
  10/26/2019 at 6:59 AM #111247
  Mauro T. “Algorithm System”

  Here’s with 200,000,
  A drawdown from a heart attack …..

  1 user thanked author for this post.

  JohnScher

  

  Mauro T. “Algorithm System”

  Participant

  Topics: 18

  Replies: 336

  Been thanked: 88 times
  10/26/2019 at 9:11 AM #111255
  GraHal

  Thank You @Mauro for the 200k results.

  Weird that your equity curve shows losses / reducing equity for most of 2018 whereas my equity curve during 2018 showed small gains?

  1 user thanked author for this post.

  JohnScher

  

  GraHal

  Participant

  Topics: 222

  Replies: 9296

  Been thanked: 2462 times
  10/27/2019 at 10:48 AM #111320
  JohnScher

  Hello and a big thanks!

  The topic that concerns me are the time-based trades due to the different behavior of the investors. If I determine this behavior I try to improve the result with simple filters. Overoptimization is to be avoided.
  Unfortunately the mentioned 200.000 K are missing to check my theory.
  So once again a big thank you to you.

  

  JohnScher

  Participant

  Topics: 45

  Replies: 338

  Been thanked: 87 times
• Author
  Posts