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.

Double Bottoms


I am trying to create a scan that would look for strong Double Bottoms. If you would please look at a chart of MPC for the dates of 5 December 2014 through 21 January 2015 that is what I am looking for. As of now the best I have is like this
[today's sctr.mid > 50] and [yesterday's max(20, sctr.mid) < 70] Thanks for your help!!

Best Answer

  • markdmarkd mod
    edited June 2015 Answer ✓
    As Kevo notes, there are problems writing scans for longer term patterns. That's not really the scan engine's fault. The market is highly variable, so there is always a trade off in how specific versus how general your search conditions should be. More general conditions are likely to return what you want, but also alot of other stuff you have to sift through. Very specific scans are more likely to get what you think you want, but often get few or no hits - and consequently miss set ups you might have accepted even if not perfect. So you have to play around with the scans to find a balance.

    The other thing is, although you are scanning for a double bottom, not every double bottom set up actually becomes a double bottom. The ratio might be 3 in 10 or whatever. What you are really scan for is, price is near where it was before prices got higher for a while. That could turn into a double bottom, or price could just keep going down from there. You have to read the chart to figure out who has the upper hand.

    For some reason, this scan does NOT pick up MPC, but it does get a few others, some of which are very good. Run it on Fridays, since it is a WEEKLY scan. Try Jan 16 2015 first to show it does get reasonable results. Also, ALL the numbers are arbitrary. You could see what happens if you eliminate some conditions.

    // begin scan
    // universe
    [group is sp500]

    // price is back in the range of the weekly bar some weeks ago
    and [weekly close < 13 weeks ago weekly high]
    and [weekly close > 13 weeks ago weekly low]

    // the bar some weeks ago was the lowest low for a few weeks before and // after
    and [13 weeks ago weekly low = 1 week ago min(21, weekly low)]

    // there was a reasonable rise in between
    and [5 weeks ago min(3, weekly low) > 13 weeks ago weekly low * 1.03]

    // the market has been coming down for a few weeks
    and [weekly close = min(7, weekly close)]

    // end scan

    By the way, sctrs won't find patterns except by accident. Sctrs compare the stock to its group, so two sets of price data are moving. When the stock come back to a previous level, it very likely will not have the same relationship to its group, which has also been moving.


Sign In or Register to comment.