Price bouncing over horizontal support on SMA 200 + Fibonacci 61.8% retracement

• 08/22/2018 at 8:31 PM #78817
  OlivierLarue

  Hi, I would like to create a screener in which the rules would be :

  • Stock which was uptrending
  • Then fell onto a horizontal price support which happen to be in line with/on the same level as :
  > SMA 200
  > Fibonacci 61.8% retracement level from previous rise (swing low to swing high point).

  The purpose is to use a combination of signals/confluence in order to find highly probable trading setups.

  How can one achieve such a screener ?

  Thank you

  

  OlivierLarue

  Participant

  Topics: 9

  Replies: 16

  Been thanked: 1 time
  08/22/2018 at 10:29 PM #78818
  OlivierLarue

  Here is the type of setup I’m trying to achieve…
  Ticker : NASDAQ:MOMO
  Date : 2016-12-19

  

  OlivierLarue

  Participant

  Topics: 9

  Replies: 16

  Been thanked: 1 time
  08/24/2018 at 8:09 AM #78891
  Nicolas

  In order to code your request, please help us to define:

  1. Stock which was uptrending
  2. swing low to swing high point (how to find these points?)

  1 user thanked author for this post.

  OlivierLarue

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  08/24/2018 at 8:48 AM #78899
  OlivierLarue

  Hi Nicolas,

   

  1) Uptrending
  We could say that the highest point (the one before the retracement) was above it’s SMA 50 and SMA 200 ?
  You may have a better idea on the subject ?

   

  2) Swing low to swing high
  I have to admit it’s a difficult one…
  I was trying to define it in simple words and even though the human brain can see at first glance the significant line on the chart, it’s not so easy to write a rule for this…

  Ideally it would be the lowest point, on the same(ish) angled line (trendline) which lead to the highest point until it retraced.
  I think it may be very difficult to write as computer code in which things are binary…

  So I was thinking maybe using the ZigZag indicator to find the lowest and highest points over a loopback period ?
  So we would just have to run the screener a few times with a different period parameter ?

  As on the first point, you may have a better idea ?

  Thank you so much for everything you have been doing on this forum 🙂

  

  OlivierLarue

  Participant

  Topics: 9

  Replies: 16

  Been thanked: 1 time
  08/24/2018 at 9:16 AM #78904
  OlivierLarue

  Here is another example on the same stock (NASDAQ:MOMO) at the time of writing (2018-08-24).

  • The 200 moving average is not exactly 61.8%, but quite close (may be the rule should be between 55% and 65% retracement ?)
  • The support is strong

  The result was a 25% rise in just 7 days…

  

  OlivierLarue

  Participant

  Topics: 9

  Replies: 16

  Been thanked: 1 time
  08/24/2018 at 10:21 AM #78920
  OlivierLarue

  As well I think it would be interesting to add that the price didn’t go below the current price since the highest point of the retracement.
  So we get the 1st rebounce on that fibonacci level.

  

  OlivierLarue

  Participant

  Topics: 9

  Replies: 16

  Been thanked: 1 time
  08/28/2018 at 9:42 AM #79233
  Nicolas

  1/ counting how many times the MA50 had been above the MA200 is a common and simple rule to determine an uptrend. How course, you’ll have to determine a threshold for the count value to confirm the trend.

  2/ storing the lowest low value while MA50 is below MA200 and the highest high while it is above could be a good idea.

  Let me know your thinking about these 2 proposals.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  08/28/2018 at 10:47 PM #79267
  OlivierLarue

  Yes Nicolas, that sounds great to me.

  

  OlivierLarue

  Participant

  Topics: 9

  Replies: 16

  Been thanked: 1 time
  08/29/2018 at 12:03 PM #79293
  Nicolas

  This first rough version is a first try and does not find the good LL and HH:

  uptrendThreshold = 20 //periods //indis ma200=average[200] ma50=average[50] //reset highest high and lowest low if ma50 crosses under ma200 then ll=low endif if ma50 crosses over ma200 then hh=0 endif //compute hh and ll hh=max(hh,high) ll=min(ll,low) //confirmed uptrend uptrend = summation[uptrendThreshold](ma50>ma200)=uptrendThreshold //retracement 61.8 percent = (close-ll)/(hh-ll) retrace = percent>=0.55 and percent<=0.65 //global test test = uptrend and retrace and ll>0 and hh>0 screener[test](percent)

  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

  uptrendThreshold = 20 //periods   //indis ma200=average[200] ma50=average[50]   //reset highest high and lowest low if ma50 crosses under ma200 then ll=low endif if ma50 crosses over ma200 then hh=0 endif   //compute hh and ll hh=max(hh,high) ll=min(ll,low)   //confirmed uptrend uptrend = summation[uptrendThreshold](ma50>ma200)=uptrendThreshold   //retracement 61.8 percent = (close-ll)/(hh-ll) retrace = percent>=0.55 and percent<=0.65   //global test test = uptrend and retrace and ll>0 and hh>0   screener[test](percent)

  My proposal for point 2 was not a good idea since moving average are obviously suffering from lag the saved LL and HH are not the ones we are looking for. I let you doing some tests and try to find a better way to find them 🙂

  I still did not add any condition on the price bouncing on MA200.

  1 user thanked author for this post.

  OlivierLarue

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  08/30/2018 at 8:39 PM #79430
  OlivierLarue

  Thank you so much Nicolas, I’ll investigate see if I can come up with some other rules 🙂

  

  OlivierLarue

  Participant

  Topics: 9

  Replies: 16

  Been thanked: 1 time
• Author
  Posts