TRYING TO MAKE POSITIONSIZE DEPENDING OF STRATEGYPROFIT

• 06/06/2018 at 11:30 AM #72452
  TempusFugit

  Hi,

  I want to make the positionsize depending of the strategyprofit over/under its average. With something like this:

  IF STRATEGYPROFIT>AVERAGE[5000](STRATEGYPROFIT) THEN POSITIONSIZE = 2 ELSE POSITIONSIZE = 1 ENDIF

  1 2 3 4 5

  IF STRATEGYPROFIT>AVERAGE[5000](STRATEGYPROFIT) THEN POSITIONSIZE = 2 ELSE POSITIONSIZE = 1 ENDIF

  It doesn´t work. Trying different ways and nothing, I am getting a little crazy about it. Anybody knows why?

   

  

  TempusFugit

  Participant

  Topics: 41

  Replies: 232

  Been thanked: 65 times
  06/06/2018 at 11:44 AM #72453
  Vonasi

  Could it be that when you start strategy profit = zero and you then comparing that to an average of zero?

  1 user thanked author for this post.

  TempusFugit

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  06/06/2018 at 11:49 AM #72455
  TempusFugit

  Maybe you are right but in that case shouldn´t it affect only the first 5000 candles (the periods of the avegarage)

  Anyway I will try to take that in consideration and will see

  Thanks Vonasi

  

  TempusFugit

  Participant

  Topics: 41

  Replies: 232

  Been thanked: 65 times
  06/06/2018 at 11:58 AM #72459
  Vonasi

  Have you seen this:

  https://www.prorealcode.com/blog/trading-strategy-profit-curve/

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  06/06/2018 at 11:59 AM #72460
  robertogozzi

  Try adding

  graph strategyprofit graph average[5000](strategyprofit)

  1 2	graph strategyprofit graph average[5000](strategyprofit)

  to your strategy, provided you have PreLoaded enough bars.

  I tried and it always shows the same value, so I guess it is not allowed, though syntax is correct!

   

  

  robertogozzi

  Moderator

  Topics: 69

  Replies: 17771

  Been thanked: 4423 times
  06/06/2018 at 12:11 PM #72463
  Vonasi

  I put this in a code and it gave me graph lines.

  sp = strategyprofit spave = Average[5000](sp) Graph sp graph spave

  1 2 3 4

  sp = strategyprofit spave = Average[5000](sp) Graph sp graph spave

   

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  06/06/2018 at 12:12 PM #72464
  TempusFugit

  Yes, I saw it before asking and I didn´t find nothing to help me. But I see now in the replays a couple of people comment that it doens´t seem to work, Nicholas says it will take it to the PRT people. This is more than two years ago, I guess it doesn´t work yet, it´s a shame it really looked like a nice way to avoid big losses with systems go bust

  I found this topic that contains other ways to get the same idea, just seemed more complicated, I liked the simplicity of the strategyprofit average 🙁

  https://www.prorealcode.com/topic/position-size-management-performance-based-increases/

   

  

  TempusFugit

  Participant

  Topics: 41

  Replies: 232

  Been thanked: 65 times
  06/06/2018 at 12:18 PM #72467
  Vonasi

  Try my code in post #72463. It appears the issue is having StrategyProfit in the average calculation. If it is a variable it works.

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  06/06/2018 at 12:20 PM #72468
  Vonasi

  So your code should be:

  sp = strategyprofit spave = Average[5000](sp) IF STRATEGYPROFIT > spave THEN POSITIONSIZE = 2 ELSE POSITIONSIZE = 1 ENDIF

  1 2 3 4 5 6 7 8

  sp = strategyprofit spave = Average[5000](sp)   IF STRATEGYPROFIT > spave THEN POSITIONSIZE = 2 ELSE POSITIONSIZE = 1 ENDIF

  Not tested yet.

  1 user thanked author for this post.

  TempusFugit

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  06/06/2018 at 12:25 PM #72469
  Vonasi

  You should be aware that StrategyProfit is only updated when positions are closed so if you have an accumulating positions strategy then if it starts with the equity curve above the average and then you keep adding to a losing position it will add positionsizes of 2 continually as it will not be aware that your floating equity curve is below the average of the equity curve.

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  06/06/2018 at 12:33 PM #72470
  TempusFugit

  I had the same idea about the variables and was trying… I had to do it a couple of times because I thought I did something wrong… no operations at all!!! It seems something is really wrong with this thing of the average/strategyprofit

  

  TempusFugit

  Participant

  Topics: 41

  Replies: 232

  Been thanked: 65 times
  06/06/2018 at 12:34 PM #72471
  TempusFugit

  No accumulated positions in this system

  

  TempusFugit

  Participant

  Topics: 41

  Replies: 232

  Been thanked: 65 times
  06/06/2018 at 12:42 PM #72472
  Vonasi

  That is strange because I just threw it into a rubbish strategy and it works for me. Graphed line is positionsize.

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  06/06/2018 at 12:45 PM #72475
  Vonasi

  …and the strategyprofit and strategyprofit average

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1812 times
  06/06/2018 at 1:00 PM #72478
  TempusFugit

  Yes, Vonasi, you are right, using your code it seems to work.

  I think I have nailed where is the key, is using a variable in the subject of the average so…

  Average[5000](strategyprofit) --> NO work Average[5000](sp) --> YES work

  1 2 3

  Average[5000](strategyprofit) --> NO work   Average[5000](sp) --> YES work

  so we don´t need the average variable, just this seems to work…

  sp = StrategyProfit IF strategyprofit > Average[5000](sp) THEN POSITIONSIZE = 2 ELSE POSITIONSIZE = 1 ENDIF

  1 2 3 4 5 6 7

  sp = StrategyProfit   IF strategyprofit > Average[5000](sp) THEN POSITIONSIZE = 2 ELSE POSITIONSIZE = 1 ENDIF

  Strange but the important thing is that the problem is solved

  Thank you

   

  1 user thanked author for this post.

  robertogozzi

  

  TempusFugit

  Participant

  Topics: 41

  Replies: 232

  Been thanked: 65 times
• Author
  Posts