Strategy Hull & Angle

• 11/18/2020 at 8:41 AM #150856
  Paul

  Here’s a lite strategy using the Hull Moving Average & the Angle based on the Hull.

  In total 2 variables and of-course a stoploss & profittarget.

  Not intended to find something perfect on 200k bars but test a few bars less, on any timeframe.

  it’s pretty good for such small code

  defparam cumulateorders = false // strategy Period= a1 inner = 2*weightedaverage[round( Period/2)](typicalprice)-weightedaverage[Period](typicalprice) hull = weightedaverage[round(sqrt(Period))](inner) cb1 = hull > hull[1] and hull[1] < hull[2] cs1 = hull < hull[1] and hull[1] > hull[2] periodea = a2 mma = average[periodea](hull) adjasuroppo = (mma-mma[1])*pipsize angle = (atan(adjasuroppo)) cb2 = angle > angle[1] and angle[1] < angle[2] cs2 = angle < angle[1] and angle[1] > angle[2] // conditions condbuy = cb1 and cb2 condsell = cs1 and cs2 // entry if condbuy then buy 1 contract at market endif if condsell then sellshort 1 contract at market endif // set stop %loss a3 set target %profit a3

  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

  defparam cumulateorders = false   // strategy Period= a1 inner = 2*weightedaverage[round( Period/2)](typicalprice)-weightedaverage[Period](typicalprice) hull = weightedaverage[round(sqrt(Period))](inner)   cb1 = hull > hull[1] and hull[1] < hull[2] cs1 = hull < hull[1] and hull[1] > hull[2]   periodea      = a2 mma = average[periodea](hull) adjasuroppo = (mma-mma[1])*pipsize angle = (atan(adjasuroppo))   cb2 = angle > angle[1] and angle[1] < angle[2] cs2 = angle < angle[1] and angle[1] > angle[2]   // conditions condbuy  = cb1 and cb2 condsell = cs1 and cs2   // entry if condbuy then buy 1 contract at market endif   if condsell then sellshort 1 contract at market endif   // set stop %loss a3 set target %profit a3

   

  6 users thanked author for this post.

  AlexMarlier, MaoRai54, josef1604, GraHal, nonetheless

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  11/18/2020 at 10:34 AM #150860
  zilliq

  Thanks @Paul

  I did a trial on EUr/USD 1 mn as often 100 k

  Results are not bad but need some improvments

  Whats is not good:

  Too much time in the market 98.41 % (Need <20 %)

  Bad Sharpe ratio : 0.4 (Need > 1.5)

  Bad Standard deviation : 1978.74 (need < average gain )

  Too much variations in day/gain

  But I think that with some change it could be good

  Hava a nice day Paul

   

  

  zilliq

  Participant

  Topics: 58

  Replies: 683

  Been thanked: 152 times
  11/18/2020 at 10:45 AM #150865
  nonetheless

  looks interesting. ran some quick tests at 5m and 15m and yes, even in this stripped down form it can turn a profit. should be useful in conjunction with other indicators, TFs etc

  thanks for that!

  1 user thanked author for this post.

  josef1604

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1235

  Been thanked: 978 times
  11/18/2020 at 3:08 PM #150872
  Paul

  Hi zilliq  I think this code is exactly as short as you like to see it! However you did mention in previous topics that you add money-management & market structure (=parameters?). Indeed the time in market is too long and also the other things you mentioned. But you say that with some change it could be good,  so now i’am curious what kind of change you are thinking off. But still nice to see positive results on eurusd!

  Hi Nonetheless If you make progress using this concept in any way, i’am always interested! I will also try to expand it.

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  11/18/2020 at 4:23 PM #150876
  zilliq

  Hi @Paul it’s short and 3 variables : Perfect 😉

  The money management can be enough with stop loss

  I think it’s possible to work on the Signal part and there is no code for the Market structure (Range/Trend) to filter the Signals

  Have a nice day Paul

  

  zilliq

  Participant

  Topics: 58

  Replies: 683

  Been thanked: 152 times
  11/18/2020 at 7:07 PM #150883
  eckaw

  This is very interesting. I have tried the same concept but my code looks different. I will take a look and post if useful.

  
  @Zilliq – what would you suggest to use for market structure? I’m curious as I have been following you on the Donkey Posts.

  

  eckaw

  Participant

  Topics: 13

  Replies: 118

  Been thanked: 61 times
  11/20/2020 at 7:06 PM #151084
  zilliq

  Hi @eckaw

  In fact you can use what you want to determine the market structure (ATR, ADX, MACD and so on…) who can determine when you are in Trend, and when your are in Range

  Always a small code (The more complex is your code, the more curefit you will do, the worst OOS results you will have), never more than 3-4 parameters (Same reaosns) and so on…Always the same advices

  And you should watch some results as on the picture on an algo on which I work actually and who is in my incubator (OOS demo since 10 days)

  Small time in market, >1.5-2 Sharpe ratio, STD gain<Average gain and so on

  Have a nice day

  

  zilliq

  Participant

  Topics: 58

  Replies: 683

  Been thanked: 152 times
  11/27/2020 at 9:04 AM #151860
  FernetTrader

  What algo is that on your last picture, @zilliq?

  

  FernetTrader

  Participant

  Topics: 0

  Replies: 1

  Been thanked: 0 times
  12/15/2020 at 11:18 AM #153824
  Tanou

  Nice strat @Paul, I’ll see if it can inspire / improve some of the strategies I am exploring 🙂

   

  @zilliq, what’s your code? 🙂

  1 user thanked author for this post.

  Paul

  

  Tanou

  Participant

  Topics: 18

  Replies: 153

  Been thanked: 29 times
  12/15/2020 at 1:47 PM #153841
  Paul

  @Tanou I doubt he will show the full code 🙂

  About this strategy, made some changes and one part uses mtf

  runs on dow 30 seconds timeframe, with adjusted trailingstop

   

   

   

   

  1 user thanked author for this post.

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  12/15/2020 at 1:55 PM #153844
  deletedaccount051022

  @Paul  You’re a legend, thank you for sharing  🙂

  1 user thanked author for this post.

  Paul

  

  deletedaccount051022

  Participant

  Topics: 26

  Replies: 138

  Been thanked: 99 times
  12/15/2020 at 2:14 PM #153849
  Tanou

  Thanks @Paul 😉

   

  However, it doesn’t looks good… A lot of transaction for a 55% win trades and 1.09 profit… :/

  

  Tanou

  Participant

  Topics: 18

  Replies: 153

  Been thanked: 29 times
  12/15/2020 at 2:19 PM #153850
  nonetheless

  Hi Paul, thanks for the new version. I see you are sticking with the old Hull coding as in v10.3

  As I’m sure you know, “hullaverage [20] close” (for example) is recognised in v11 but it gives completely different results. I can’t say if it’s better or worse, but worth knowing that the 2 are not interchangeable.

  I’m staying with the old-style code as well, as that’s the one I know.

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1235

  Been thanked: 978 times
  12/15/2020 at 2:45 PM #153853
  Paul

  @Tanou

  that’s no good indeed!

  It’s optimised on 100k or 200k I thought. But if loading 1 million it’s bad. Time is too short optimising such amount. The way I was trying find 1 or 2 values out of 4 with normal optimisation on say 100k and set them fixed. WF the other ones. Still it’s difficult to have a good wf results oos.

  But I tried such approach in demo after good wf results and it’s positive, but sometimes a max loss can take progress away, therefore the changed trailing stop but I haven’t run that.

  
  @nonetheless Hi, thanks for the reminder. I knew but forgot that completely! Will test too to see the differences.

   

   

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
• Author
  Posts