State Space

• 05/30/2016 at 8:43 PM #8432
  Nicolas

  State space reminds me of brownian motion.

  They are no magic behind the results the optimizer give you, it’s just optimization of your 7 variables I believe. But it makes me curious so I did an indicator with the best optimized values of these 7 variables. As far as I can see the return value is so high that PRT cannot give its real because it is out of scale.. This is very weird..

  This is the indicator code:

  d1=0 d2=0 d3=1 d4=-1 n1=1 n2=1 n3=1 n4=0 dx1=-d1*x1-d2*x2-d3*x3-d4*x4+close dx2=x1 dx3=x2 dx4=x3 x1=x1+dx1 x2=x2+dx2 x3=x3+dx3 x4=x4+dx4 StateSpace=n1*x1+n2*x2+n3*x3+n4*x4 RETURN statespace

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

  d1=0 d2=0 d3=1 d4=-1   n1=1 n2=1 n3=1 n4=0   dx1=-d1*x1-d2*x2-d3*x3-d4*x4+close dx2=x1 dx3=x2 dx4=x3   x1=x1+dx1 x2=x2+dx2 x3=x3+dx3 x4=x4+dx4   StateSpace=n1*x1+n2*x2+n3*x3+n4*x4   RETURN statespace

  .. but as you can see in the attached pictures, the “statespace” variables give almost dumb values which tends to signify that optimize it is like random coin toss.

  Sorry, but I don’t think that there is something very interesting here.

   

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4774 times
  05/31/2016 at 6:19 AM #8454
  simon_json

  Thanks Nicolas! Please test it on daily TF. Do you think it possibile to get the exact same result on 5min TF by changing some code (on the same period of time, one year for example).

  

  simon_json

  Participant

  Topics: 14

  Replies: 53

  Been thanked: 5 times
  05/31/2016 at 8:02 AM #8460
  Nicolas

  As I told you, there is nothing here. So you can optimize on every timeframe you’d like, you can have good results, like almost any kind of strategy. Sorry.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4774 times
• Author
  Posts