New Members: Be sure to confirm your email address by clicking on the link that was sent to your email inbox. You will not be able to post messages until you click that link.

Help creating a scan to pick up 'cancellation' candlestick reversal patterns please

Best current example I can provide GE (General Electric) Monthly chart, spinning top closed lower March 2018, then spinning top closed higher April 2018 (highs and lows, open and closes of each month were almost identical)... I can it 'cancellation' pattern but would appreciate any help creating a scan that can pick up these types of patterns.
Thanks in advance!


  • You would have to determine how close to identical you want the two corresponding prices to be. Let's say it's 1 per cent. Then you would write

    and [monthly high < 1 month ago monthly high * 1.01]
    and [monthly high > 1 month ago monthly high * .99]

    Repeat for all four prices.

    Following the GE example, if you want this combination to occur at the end of a trend, you would have to specify that, too. So you would compare the values of a moving average for at least two points in time. So,

    and [ monthly sma(12, monthly close) < 6 months ago monthly sma(12, monthly close)]

    You could extend that with another comparison, say

    and [ 6 months ago monthly sma(12, monthly close) < 12 months ago monthly sma(12, monthly close)]

    and so on.

    You would have to play with all the parameters to see what works better. I'm not sure how common these patterns are, so it could get frustrating. Also, when you do find them, the follow through may not be what you expect - in other words, in your GE example, the pattern looks like a potential reversal. But in your research you may find that you get hits where the trend continues in its current direction.
  • Thank you very much for this info Mark D - great start for sure!

    Yes, other research goes into this (professional options trader) so only need to hit short price targets; not banking on a turnaround :)

    I've added a few extra screeners, really tightened up the % and adapted it for weekly - its given me some great leads, but still some really wonky outputs: MTCH comes up for example??

    [type is stock] and [sma(20,volume) > 1000000] and [Close > 5] and [region is NorthAmerica]
    and [weekly high < 1 week ago weekly high * 1.001]
    and [weekly high > 1 week ago weekly high * .999]

    Let me know your thoughts.
  • Well, to avoid things like MTCH, you need to expand the scan to make similar comparisons for the weekly low, the weekly close and weekly open.
Sign In or Register to comment.