% trail with acceleration

• 02/27/2024 at 1:55 PM #228833
  nonetheless

  hi, this is the %trail i’ve been using for some time. it’s mostly Paul’s work with some minor alterations of my own:

  // %trailing stop (inc. cumulative) once trailingstoptype = 1 if trailingstoptype then //==================== trailingpercentlong = tsl // % trailingpercentshort = tss // % once acceleratorlong = acl // typically tsl*0.1 once acceleratorshort= acs // typically tss*0.1 ts2sensitivity = 2 // [1] close [2] high/low [3] low/high [4] typicalprice PcToKeep = pk //long % to keep at breakeven PcToKeepS = pks //short % to keep at breakeven //==================== once steppercentlong = (trailingpercentlong/10)*acceleratorlong once steppercentshort = (trailingpercentshort/10)*acceleratorshort if onmarket then trailingstartlong = positionprice*(trailingpercentlong/100) trailingstartshort = positionprice*(trailingpercentshort/100) trailingsteplong = positionprice*(steppercentlong/100) trailingstepshort = positionprice*(steppercentshort/100) endif if not onmarket or ((longonmarket and shortonmarket[1]) or (longonmarket[1] and shortonmarket)) then newsl = 0 mypositionprice = 0 endif positioncount = abs(countofposition) if newsl > 0 then if positioncount > positioncount[1] then if longonmarket then newsl = max(newsl,positionprice * newsl / mypositionprice) else newsl = min(newsl,positionprice * newsl / mypositionprice) endif endif endif if ts2sensitivity=1 then ts2sensitivitylong=close ts2sensitivityshort=close elsif ts2sensitivity=2 then ts2sensitivitylong=high ts2sensitivityshort=low elsif ts2sensitivity=3 then ts2sensitivitylong=low ts2sensitivityshort=high elsif ts2sensitivity=4 then ts2sensitivitylong=(typicalprice) ts2sensitivityshort=(typicalprice) endif if longonmarket then if newsl=0 and ts2sensitivitylong-positionprice>=trailingstartlong then newsl = positionprice+trailingsteplong+(positionprice/100*pk) endif if newsl>0 and ts2sensitivitylong-newsl>=trailingsteplong then newsl = newsl+trailingsteplong endif endif if shortonmarket then if newsl=0 and positionprice-ts2sensitivityshort>=trailingstartshort then newsl = positionprice-trailingstepshort-(positionprice/100*pks) endif if newsl>0 and newsl-ts2sensitivityshort>=trailingstepshort then newsl = newsl-trailingstepshort endif endif if barindex-tradeindex>1 then if longonmarket then if newsl>0 then sell at newsl stop endif if newsl>0 then if low crosses under newsl then sell at market endif endif endif if shortonmarket then if newsl>0 then exitshort at newsl stop endif if newsl>0 then if high crosses over newsl then exitshort at market endif endif endif endif mypositionprice = positionprice endif if (shortonmarket and newsl > 0) or (longonmarket and newsl>0) then if positioncount > positioncount[1] then if longonmarket then newsl = max(newsl,positionprice * newsl / mypositionprice) endif if shortonmarket then newsl = min(newsl,positionprice * newsl / mypositionprice) endif endif endif

  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 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99

  // %trailing stop (inc. cumulative) once trailingstoptype = 1 if trailingstoptype then //==================== trailingpercentlong  = tsl // % trailingpercentshort = tss // % once acceleratorlong = acl // typically tsl*0.1 once acceleratorshort= acs // typically tss*0.1 ts2sensitivity  = 2 // [1] close [2] high/low [3] low/high [4] typicalprice PcToKeep = pk   //long % to keep at breakeven PcToKeepS = pks //short % to keep at breakeven //==================== once steppercentlong  = (trailingpercentlong/10)*acceleratorlong once steppercentshort = (trailingpercentshort/10)*acceleratorshort if onmarket then trailingstartlong = positionprice*(trailingpercentlong/100) trailingstartshort = positionprice*(trailingpercentshort/100) trailingsteplong = positionprice*(steppercentlong/100) trailingstepshort = positionprice*(steppercentshort/100) endif if not onmarket or ((longonmarket and shortonmarket[1]) or (longonmarket[1] and shortonmarket)) then newsl           = 0 mypositionprice = 0 endif positioncount = abs(countofposition) if newsl > 0 then if positioncount > positioncount[1] then if longonmarket then newsl = max(newsl,positionprice * newsl / mypositionprice) else newsl = min(newsl,positionprice * newsl / mypositionprice) endif endif endif if ts2sensitivity=1 then ts2sensitivitylong=close ts2sensitivityshort=close elsif ts2sensitivity=2 then ts2sensitivitylong=high ts2sensitivityshort=low elsif ts2sensitivity=3 then ts2sensitivitylong=low ts2sensitivityshort=high elsif ts2sensitivity=4 then ts2sensitivitylong=(typicalprice) ts2sensitivityshort=(typicalprice) endif if longonmarket then if newsl=0 and ts2sensitivitylong-positionprice>=trailingstartlong then newsl = positionprice+trailingsteplong+(positionprice/100*pk) endif if newsl>0 and ts2sensitivitylong-newsl>=trailingsteplong then newsl = newsl+trailingsteplong endif endif if shortonmarket then if newsl=0 and positionprice-ts2sensitivityshort>=trailingstartshort then newsl = positionprice-trailingstepshort-(positionprice/100*pks) endif if newsl>0 and newsl-ts2sensitivityshort>=trailingstepshort then newsl = newsl-trailingstepshort endif endif if barindex-tradeindex>1 then if longonmarket then if newsl>0 then sell at newsl stop endif if newsl>0 then if low crosses under newsl then sell at market endif endif endif if shortonmarket then if newsl>0 then exitshort at newsl stop endif if newsl>0 then if high crosses over newsl then exitshort at market endif endif endif endif mypositionprice = positionprice endif   if (shortonmarket and newsl > 0) or (longonmarket and newsl>0) then if positioncount > positioncount[1] then if longonmarket then newsl = max(newsl,positionprice * newsl / mypositionprice) endif if shortonmarket then newsl = min(newsl,positionprice * newsl / mypositionprice) endif endif endif

  then i had the idea that the accelerator could be increased under certain circumstances, so i tried changing lines 13,14 to this:

  steppercentlong = (trailingpercentlong/10)*acceleratorlong if (longonmarket and cs6) then steppercentlong = (trailingpercentlong/10)*(acceleratorlong*s1) endif steppercentshort = (trailingpercentshort/10)*acceleratorshort if (shortonmarket and cb6) then steppercentshort = (trailingpercentshort/10)*(acceleratorshort*s2) endif

  1 2 3 4 5 6 7 8

  steppercentlong  = (trailingpercentlong/10)*acceleratorlong if (longonmarket and cs6) then steppercentlong  = (trailingpercentlong/10)*(acceleratorlong*s1) endif steppercentshort = (trailingpercentshort/10)*acceleratorshort if (shortonmarket and cb6) then steppercentshort  = (trailingpercentshort/10)*(acceleratorshort*s2) endif

  in my test, the algo was running at 5m and cs6/cb6 represent a change of direction for the stochRSI on the 15m TF; when that starts going the wrong way, the accelerator increases.

  s1 and s2 are the multipliers, how much faster you want the step to move.

  this seems to give a small advantage, esp for shorts, but i was expecting more; is there a better way of doing it?

  thanks in advance for any suggestions.

  1 user thanked author for this post.

  robertogozzi

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  02/27/2024 at 3:08 PM #228839
  PeterSt

  I have so many variations on this, that I am pretty sure that no one rule can exist for it. Combine it with where trailing actually starts (see below) and the appliances are endless. Or the trailing distance itself (2nd set you see below for close-by).
  One thing I personally know for sure : also influencing the slope underway (your subject) is only creating extra complexity in your code. All it does further is giving the opportunity to optimize over it. Something I handed out the death penalty over to myself.

  With trailing you’ll earn half of the money. But optimizing over it (any of the trailing parameters) is only not-good-habit. You can just as well set it for once by visualizing it like I show below (that comprises 35 or so pictures for working out on of the parameters) which makes you dizzy anyway, and next let it be waiting for the best” trails (always very nice to watch).

  This mere vague post should try to bring across in the base, that – for me – no changing slopes work (considered slopes which even can turn negative already ?). It is just something without control, BUT which may run into the price at a lucky moment (steeper slope) or avoids the price which gets better later (shallower slope). Make it a linear slope and you do the same, but now with control (see the trend, follow it).

  Helpful a little ? I hope so.
  The best part : your posting in itself. 🙂 🙂

   

   

   

  1 user thanked author for this post.

  nonetheless

  

  PeterSt

  Participant

  Topics: 70

  Replies: 2287

  Been thanked: 695 times
  02/27/2024 at 3:27 PM #228843
  nonetheless

  hi Peter, thanks for that, an interesting take as always – hope all is good with you!

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
• Author
  Posts