Double Top bottom

• 05/18/2023 at 12:03 AM #214867
  Aragorna

  fractalP = 10 //fractal period percent = 0.030 //maximum percent change between the 2 points barlimit = 100 //maximum bars between 2 points //fractals cp=fractalP if high[cp] >= highest[(cp)*2+1](high) then //new fractal high found $TOPy[lastset($TOPy)+1] = high[cp] //store fractal value $TOPx[lastset($TOPx)+1] = barindex[cp] //store fractal barindex //COND=Summation[fractalP](Close>$topy[i])=1 endif if low[cp] <= lowest[(cp)*2+1](low) then //new fractal low found $BOTy[lastset($BOTy)+1] = low[cp] //store fractal value $BOTx[lastset($BOTx)+1] = barindex[cp] //stire fractal barindex endif if(islastbarupdate and isset($topy[0]) and isset($boty[0])) then //check points in a range of X percent for i = 0 to lastset($TOPy) do //loop through the tops for y = 0 to lastset($TOPy) do //check first top with other tops change=abs(($topy[y]-$topy[i])/$topy[i]) //percent range between the 2 tops bardiff=abs($TOPx[i]-$TOPx[y]) //how many bars between the 2 points? Dtop= change<=percent/100 and bardiff<barlimit and bardiff>=fractalP and $topx[i]<>$topx[y] next next for i = 0 to lastset($BOTy) do //loop through the bottoms for y = 0 to lastset($BOTy) do //check first bottom with other bottoms change=abs(($boty[y]-$boty[i])/$boty[i]) //percent range between the 2 bottoms bardiff=abs($botx[i]-$botx[y]) //how many bars between the 2 points? Dbot= change<=percent/100 and bardiff<barlimit and bardiff>=fractalP and $BOTx[i]<>$BOTx[y] next next endif if not LONGONMARKET and Dbot then buy 1 shares AT MARKET ENDIF Set Stop loss 30 Set Target profit 60

  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 32 33 34 35 36 37 38 39 40 41

  fractalP = 10 //fractal period percent = 0.030 //maximum percent change between the 2 points barlimit = 100 //maximum bars between 2 points   //fractals cp=fractalP if high[cp] >= highest[(cp)*2+1](high) then //new fractal high found $TOPy[lastset($TOPy)+1] = high[cp] //store fractal value $TOPx[lastset($TOPx)+1] = barindex[cp] //store fractal barindex //COND=Summation[fractalP](Close>$topy[i])=1 endif   if low[cp] <= lowest[(cp)*2+1](low)  then //new fractal low found $BOTy[lastset($BOTy)+1] = low[cp] //store fractal value $BOTx[lastset($BOTx)+1] = barindex[cp] //stire fractal barindex endif   if(islastbarupdate and isset($topy[0]) and isset($boty[0])) then //check points in a range of X percent for i = 0 to lastset($TOPy) do //loop through the tops for y = 0 to lastset($TOPy) do //check first top with other tops change=abs(($topy[y]-$topy[i])/$topy[i]) //percent range between the 2 tops bardiff=abs($TOPx[i]-$TOPx[y]) //how many bars between the 2 points? Dtop= change<=percent/100 and bardiff<barlimit and bardiff>=fractalP and $topx[i]<>$topx[y] next next   for i = 0 to lastset($BOTy) do //loop through the bottoms for y = 0 to lastset($BOTy) do //check first bottom with other bottoms change=abs(($boty[y]-$boty[i])/$boty[i]) //percent range between the 2 bottoms bardiff=abs($botx[i]-$botx[y]) //how many bars between the 2 points? Dbot= change<=percent/100 and bardiff<barlimit and bardiff>=fractalP and $BOTx[i]<>$BOTx[y] next next endif   if not LONGONMARKET and Dbot then buy 1 shares AT MARKET ENDIF Set Stop loss 30 Set Target profit 60

  hello to all, I’ve got a simple request. I’m studying the double top-bottom made by Nicolas, I’m trying to make a strategy, just for study. I’d like to go long when  the second bottom at the condition Dbot) is identified, and go short when the second top is identifed. Strategy I’ve tryed does not give any result. can someone give me a tip? thank’0s in advance

  Alessio

  

  Aragorna

  Participant

  Topics: 37

  Replies: 130

  Been thanked: 12 times
  05/18/2023 at 10:49 AM #214876
  GraHal

  Aragorna wrote:

  Strategy I’ve tryed does not give any result

  No result in backtest or in Live trading?

  What Instrument, TF are you testing on?

  

  GraHal

  Participant

  Topics: 214

  Replies: 8880

  Been thanked: 2222 times
  05/18/2023 at 11:37 AM #214879
  GraHal

  I can comfirm that the above code does run on Daily TF on DJI, but results in zero trades.

  Please can you provide a link to the source of the above code? I have searched the Library with no success.

  

  GraHal

  Participant

  Topics: 214

  Replies: 8880

  Been thanked: 2222 times
  05/18/2023 at 12:05 PM #214884
  JS

  Hi GraHal,

  It’s example 6 from the array examples…
  This indicator was not set up as a trading system but perhaps it is possible to turn it into a trading system…

  1 user thanked author for this post.

  GraHal

  

  JS

  Participant

  Topics: 27

  Replies: 1814

  Been thanked: 884 times
  05/18/2023 at 12:56 PM #214885
  Aragorna

  Hi Grahal, did you find it? code is a little modified but it’s the same in the end. if you load the indicator, you can set it as it’s set in the strategy. I cannot explain why that condition does not work, I’ve graphed it and condition never shows.

  

  Aragorna

  Participant

  Topics: 37

  Replies: 130

  Been thanked: 12 times
  05/18/2023 at 1:15 PM #214887
  GraHal

  Aragorna wrote:

  I’ve graphed it and condition never shows.

  My conclusion (maybe wrongly?) was that the coincident conditions (x AND y AND z etc) do not occur coincident with each other?

  Also it is frustrating Algo to play with (for me) as, on less than Daily TF, it takes ages to run / optimise etc?

  

  GraHal

  Participant

  Topics: 214

  Replies: 8880

  Been thanked: 2222 times
  05/18/2023 at 1:32 PM #214889
  Aragorna

  conditions occur in the indicator, they have to occur in the strategy as well, I guess. you can test with less candles.

   

  

  Aragorna

  Participant

  Topics: 37

  Replies: 130

  Been thanked: 12 times
  05/18/2023 at 2:08 PM #214891
  Aragorna

  maybe Nicolas or Robertogozzi could help us?

   

  

  Aragorna

  Participant

  Topics: 37

  Replies: 130

  Been thanked: 12 times
  05/18/2023 at 2:14 PM #214892
  GraHal

  Aragorna wrote:

  I guess. you can test with less candles.

  I was down to 1k bars, but I still lost interest as it took so long … due to the arrays I would think?

  

  GraHal

  Participant

  Topics: 214

  Replies: 8880

  Been thanked: 2222 times
  05/18/2023 at 3:03 PM #214894
  Aragorna

  it doesn’t take so long for me, perhaps your personal computer isn’t powerful enough?

  

  Aragorna

  Participant

  Topics: 37

  Replies: 130

  Been thanked: 12 times
  05/18/2023 at 4:45 PM #214895
  GraHal

  It was quick on Daily but no results so I was trying 1 hour and 5 mins etc and optimising the variables … then it seemd like it had stalled! 🙂

  Optimisation is executed remotely on the PRT servers, our PC do very little so we don’t need super fast PC.

  

  GraHal

  Participant

  Topics: 214

  Replies: 8880

  Been thanked: 2222 times
  05/22/2023 at 2:07 PM #215015
  Aragorna

  It’s a pity that no one can help to understand. It could have a big potential.

  

  Aragorna

  Participant

  Topics: 37

  Replies: 130

  Been thanked: 12 times
  05/22/2023 at 2:46 PM #215019
  GraHal

  Do you understand arrays enough to manipulate the code for better results? Are you able to state exactly what you want to be changed in the code you posted?

  I think there are very few on here who can code arrays successfully?

  Maybe choose a simpler Algo and start a separate Thread?

  

  GraHal

  Participant

  Topics: 214

  Replies: 8880

  Been thanked: 2222 times
  05/22/2023 at 3:41 PM #215024
  PeterSt

  FWIW :

  I tried the original code (it is an Indicator), and this works (as indicator indeed). But there it stops because it is wayyyyyy too slow to be workable.
  Copying the code into the program itself does not help, except for that you can see that it won’t trade indeed. This is most probably because the trade points – as shown by the indicator – don’t fall on the bar-ends (the close of them)  so while there would exist a moment that this coincides, in practice you can’t find his because 1000 bars already take 15 minutes to backtest over here and in these 1K bars (of 1 minute) the situation does not occur or it is too tedious to try a few things.

  The solution, in my view, would be to enumerate the arrays one time per backtest (Optimization Parameters combination), so all the array-filling work can be done one time only, instead of 1000 times for testing over 1000 bars (or 200,000 times for 200K bars). But there it stopped to be funny, and I quit this case.

  I did not make screenshots, but they would tell that for the intention of this indicator – trade on the 2nd top etc. – it requires refinement. Read : there would be too many false alarms because the shapes / fractals are not always caught correctly. On the other hand, notice that Nicolas  just used this for a rough “arrays” example, which was a nice job to begin with.

  Now if someone could cause that arrays are being dealt with smarter by PRT (like 10,000 times more efficient), “we” could all try to work on the improvement of such a really working DT/DB indicator. This one really has a good base (says me 😉 ).

  1 user thanked author for this post.

  GraHal

  

  PeterSt

  Participant

  Topics: 73

  Replies: 2555

  Been thanked: 837 times
  05/22/2023 at 4:22 PM #215026
  Aragorna

  I understand how to manipulate arrays and also create new ones in other indicators. What I do not understand is how to use in a strategy because what I think is correct in conditions of buy and sell does not  work

  

  Aragorna

  Participant

  Topics: 37

  Replies: 130

  Been thanked: 12 times
• Author
  Posts