Validate the efficiency of the system before position

• 08/24/2020 at 3:59 PM #142448
  winnie37

  I would like to add a very simple condition to my systems, based only on observation.

  When the profit curve goes below its moving average (to be defined), positions are not taken in real life. This is a sign that the system is no longer efficient enough.

  When the profit curve drops below its moving average, positions are taken. This is a sign that the system is working well.

  This condition could apply to any system.
  Help to code these additionnal conditions?

  For example, positions on this system would not have been taken in july (profit under the moving average)

  

  winnie37

  Participant

  Topics: 30

  Replies: 167

  Been thanked: 16 times
  08/24/2020 at 4:18 PM #142450
  Vonasi

  Perhaps you have not read this blog by Nicolas yet?

  https://www.prorealcode.com/blog/learning/how-to-improve-a-strategy-with-simulated-trades-1/

  We are all still waiting for Part 2 because Nicolas is a very busy person!

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  08/24/2020 at 4:37 PM #142451
  winnie37

  Yes, I had indeed read it. The idea dates back to 2018. That’s why I’m relaunching the subject to our talented coders 🙂
  I currently test it manually, not very pratical but for testing.

  

  winnie37

  Participant

  Topics: 30

  Replies: 167

  Been thanked: 16 times
  08/24/2020 at 5:30 PM #142456
  Nicolas

  The moving average is calculated upon orders, not time. How do you calculate the MA if you don’t want to take order? The solution is to take fake orders to simulate the strategy in order to see it performs well again…

  

  Nicolas

  Keymaster

  Topics: 83

  Replies: 20898

  Been thanked: 4866 times
  08/24/2020 at 6:10 PM #142461
  Vonasi

  Perhaps this topic from when I was playing with ideas on trading the equity curve is also of interest:

  https://www.prorealcode.com/topic/simulated-trading/

  1 user thanked author for this post.

  winnie37

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  08/25/2020 at 11:16 AM #142500
  winnie37

  that’s the whole problem. Currently, I do a weekly backtest. And when the moving average goes under the profit basket, I manually stop the live system for the week.
  And I check again the following week for a restart or not of the system …

  Not very practical and precise, I admit.

  

  winnie37

  Participant

  Topics: 30

  Replies: 167

  Been thanked: 16 times
  08/25/2020 at 11:56 AM #142502
  MAKSIDE

  winnie37 wrote:

  I would like to add a very simple condition to my systems, based only on observation.

  When the profit curve goes below its moving average (to be defined), positions are not taken in real life. This is a sign that the system is no longer efficient enough.

  When the profit curve drops below its moving average, positions are taken. This is a sign that the system is working well.

  This condition could apply to any system.

  Help to code these additionnal conditions?

  For example, positions on this system would not have been taken in july (profit under the moving average)

   

  I saw that you use Day Dax of Doctrading.
  The system isn’t efficient because it’s over optimized..  (it’s my point of view..)

  It’s the same with Day DJI or Cac..  everything is wonderful until bots are put in production.

   

  1 user thanked author for this post.

  GraHal

  

  MAKSIDE

  Participant

  Topics: 38

  Replies: 345

  Been thanked: 99 times
  08/25/2020 at 12:52 PM #142510
  Nicolas

  BTW, with arrays, simulated orders for equity curve trading would be easier to code now. Need v11 and an example of code now 😉

  1 user thanked author for this post.

  winnie37

  

  Nicolas

  Keymaster

  Topics: 83

  Replies: 20898

  Been thanked: 4866 times
• Author
  Posts