Mother of Dragons trading strategy…

• 05/21/2020 at 8:51 PM #132695
  Paul

  hi nonetheless

  I was curious how your code with reentry would perform. Didn’t optimise. Uses latest atr be/ts stops from vectorial dax.

   

   

   

  4 users thanked author for this post.

  nonetheless, Scooby, Sebastian Arsjo, Francesco

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/21/2020 at 9:01 PM #132698
  Francesco

  This is very interesting, maybe could be worth apply the reentry also on the dax version. I’ll make some tests

  2 users thanked author for this post.

  Paul, Scooby

  

  Francesco

  Participant

  Topics: 32

  Replies: 439

  Been thanked: 200 times
  05/21/2020 at 10:33 PM #132722
  nonetheless

  Paul wrote:

  how your code with reentry would perform

  Thanks very much for that, not something I would have thought to do. I’ll have a closer look at it tomorrow. And as Francesco says, could also apply to the DAX which  has really been the best performer in recent months.

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  05/22/2020 at 1:37 PM #132815
  Paul

  works for the dax too I think. How did your tests go Francesco?

   

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/22/2020 at 3:05 PM #132840
  nonetheless

  Optimisation gets an improvement but not quite up to the original. This is odd  because the ATR trail works better almost everywhere else I’ve tried it, and the reenter should help. I tried a compromise of conventional break even with ATR trail but was worse. Attached image is with these values/changes:

  //if dayofweek<5 then ctime= hour>1 //elsif dayofweek=5 then //ctime= hour>1 and hour<t2 //endif // breakeven stop atr once breakevenstoptype = 1 // breakeven stop - 0 off, 1 on once beatrperiod = 10 // atr parameter value once beminstop = 12 // minimum breakeven stop distance once bepointsorperct= 1 //[0]percentage [1]points if tradetype=1 then once bestoplong = 5 // ts atr distance once bestopshort = 5 // ts atr distance bepointstokeep = 3 //-(sl/2) // positive or negative //not use once once trailingstoptype = 1 // trailing stop - 0 off, 1 on once tsatrperiod = 10 // ts atr parameter once tsminstop = 12 // ts minimum stop distance if tradetype=1 then once tsincrements = .25 // set to 0 to ignore tsincrements once tsminatrdist = 1 once tssensitivity = 0 // [0]close;[1]high/low elsif (tradetype=2 or tradetype=3) then once tsincrements = .25 // set to 0 to ignore tsincrements once tsminatrdist = 1 once tssensitivity = 1 // [0]close;[1]high/low endif if trailingstoptype then if barindex=tradeindex then if tradetype=1 then trailingstoplong = 8 // ts atr distance trailingstopshort = 8 // ts atr distance elsif (tradetype=2 or tradetype=3) then trailingstoplong = 6 // ts atr distance trailingstopshort = 6 // ts atr distance 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

  //if dayofweek<5 then ctime= hour>1 //elsif dayofweek=5 then //ctime= hour>1 and hour<t2 //endif           // breakeven stop atr once breakevenstoptype     = 1    // breakeven stop - 0 off, 1 on   once beatrperiod    = 10    // atr parameter value once beminstop      = 12    // minimum breakeven stop distance   once bepointsorperct= 1     //[0]percentage [1]points   if tradetype=1 then once bestoplong     = 5   // ts atr distance once bestopshort    = 5   // ts atr distance bepointstokeep      = 3  //-(sl/2) // positive or negative //not use once               once trailingstoptype     = 1    // trailing stop - 0 off, 1 on   once tsatrperiod    = 10         // ts atr parameter once tsminstop      = 12         // ts minimum stop distance   if tradetype=1 then once tsincrements = .25         // set to 0 to ignore tsincrements once tsminatrdist = 1 once tssensitivity        = 0    // [0]close;[1]high/low elsif (tradetype=2 or tradetype=3) then once tsincrements = .25         // set to 0 to ignore tsincrements once tsminatrdist = 1 once tssensitivity        = 1    // [0]close;[1]high/low endif   if trailingstoptype then if barindex=tradeindex then if tradetype=1 then trailingstoplong     = 8   // ts atr distance trailingstopshort    = 8   // ts atr distance elsif (tradetype=2 or tradetype=3) then trailingstoplong     = 6   // ts atr distance trailingstopshort    = 6   // ts atr distance endif

  Any other changes worth trying? Fifi’s pivots are inactive i think but I’m not sure how to bring them into play

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  05/22/2020 at 3:23 PM #132845
  Francesco

  Paul wrote:

  works for the dax too I think. How did your tests go Francesco?

  I’m going to take a break at the sea this weekend, i’ll put my hands on it next week 🙂

  2 users thanked author for this post.

  Paul, Moda

  

  Francesco

  Participant

  Topics: 32

  Replies: 439

  Been thanked: 200 times
  05/22/2020 at 9:56 PM #132899
  Franro

  Just a question,, Are you keeping the bots open or closing all positions before weekend?

  

  Franro

  Participant

  Topics: 25

  Replies: 219

  Been thanked: 23 times
  05/23/2020 at 3:32 PM #132953
  nonetheless

  I leave positions open over the weekend. I know that lots of people close everything friday evening for fear of an adverse gap when they open again, but my tests tend to show its better to let them run. Over a length of time it’s just as likely to gap in your favour.

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  05/23/2020 at 4:56 PM #132963
  Sebastian Arsjo

  I close everything before the weekend . Maybe not as effective but i like to be “free”over the weekend no matter what goes on in the world

  I like to clear my mind and start new every week .

  1 user thanked author for this post.

  swedshare

  

  Sebastian Arsjo

  Participant

  Topics: 1

  Replies: 14

  Been thanked: 4 times
  05/23/2020 at 6:18 PM #132972
  Paul

  nonetheless wrote:

  Any other changes worth trying?

  not really a.t.m. The goal of reentry is to see quickly how stable the strategy is and it looks pretty good!

   

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/23/2020 at 10:10 PM #132993
  nonetheless

  I think I misunderstood, is reentry a test rather than a part of the strategy? Something like vrt?

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  05/24/2020 at 1:32 AM #133001
  Paul

  yes a bit

  often when in a position and the same entryconditions are true again, you don’t act on those. The bigger the exit criteria are (i.e. sl/pt/ts) the more signals you could have in the same direction and it’s those signals i’am interested in .

  If taken version dow 1p short, set tradetype at 3, reenter at 1 and positionperftype at 0,1 or 2. Equitycurve & stats looks impressive & solid!  

  for long only it shows that signals are not as good as the shorts. Here is a bigger difference depending on positionperftype for the equitycurve & stats.

  With reentry, the new entry is likely to have a worse price & there are extra spread costs, which makes it the more impressive if the equity curve can hold up.

  I see it as an extra tool to create more robust parameters with optimisation and is something to experiment with.

  2 users thanked author for this post.

  nonetheless, Sebastian Arsjo

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/24/2020 at 1:33 AM #133002
  Paul

  at the bottom of the dow version you can add this

  if ctime then if (tradetype=1 or tradetype=2) and condbuy then buysignal=1 else buysignal=0 endif if (tradetype=1 or tradetype=3) and condsell then sellsignal=-1 else sellsignal=0 endif endif graph buysignal coloured(121,141,35,255) graph sellsignal coloured(255,0,0,255)

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

  if ctime then if (tradetype=1 or tradetype=2) and condbuy then buysignal=1 else buysignal=0 endif if (tradetype=1 or tradetype=3) and condsell then sellsignal=-1 else sellsignal=0 endif endif   graph buysignal coloured(121,141,35,255) graph sellsignal coloured(255,0,0,255)

   

  1 user thanked author for this post.

  swedshare

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/26/2020 at 4:54 AM #133268
  Paul

  congrats nonetheless

  it’s working nicely!

  1 user thanked author for this post.

  nonetheless

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  05/26/2020 at 7:43 AM #133275
  nonetheless

  Hey, thanks for the vote of confidence. TBH I’m amazed that it works at all, but so far not too bad. Reentry is an interesting tool, I haven’t had time to play around with it yet.

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
• Author
  Posts