bert123
53 posts
msg #71648
- Ignore bert123 |
2/22/2009 2:36:40 PM
Is there a statistically significant candlestick pattern that precedes a Gap Up? Can a Stats Filter be designed that will show
which pattern is most statistically significant comparing how many times a particular pattern was evident against the percentage of times that pattern resulted in a gap up? Thanks in advance to the coders who can figure this out!
|
chetron
2,817 posts
msg #71650
- Ignore chetron |
2/22/2009 3:39:15 PM
SORRY, BERT, THERE IS NO PATTERN IN EXISTENCE THAT PRECIDES A GAP.
|
decipherlinda
133 posts
msg #71663
- Ignore decipherlinda |
2/23/2009 8:30:39 AM
SF isn't set up to do this but you could do it manually.
Bert123, you may already know this but to get a general feel, you may want to...
Define the type of gap you want, generally: */general idea only, don't have time to check the code*/
open > yesterday's close
and open > yesterday's open
and close > yesterday's close
and close > yesterday's open (you could add the %up you want the gap and whether you want a white candle gapping up)
Run the filter, record the count, then add more criteria to the code such as "and close yesterday > open yesterday" to see if yesterday's candle was white, record the count, make manual % calculations. You'd probably also want to check which candlestick patterns were in play the day before the gap up such as the morningstar doji.
This can only be done at the end of the day after the markets have settled so the number of stocks gapping up today doesn't change minute-by-minute.
The resulting statistics show which candles are most associated with gaps up but not which are causative. I have noticed from a very quick look, that they tend to be small white or black candles, a possibly surprising number of doji's, and some long white candles.
The Japanese have studied this for a couple hundred years. If there was a good association, I would think you'd find it easier through more advanced studies of candlestick charting with Steve Nison and others.
Even so, just because there's a famous saying "Don't reinvent the wheel", doesn't mean the direction you're interested in won't work or at least help build a gut instinct!
Good luck!
|
miketranz
961 posts
msg #71681
- Ignore miketranz |
2/23/2009 9:11:13 PM
Bert,if you can't get the answer from a filter,do your own research.If you're looking to climb aboard a stock getting ready to gap from a candlestick pattern,I think you're looking for a needle in a hay stack.You might want to check out some of the gap filters on here.Even though they 're not exactly what your looking for,they may have something to offer.I myself play overnight gaps,buy on close,sell on open.Certain stocks have a high percentage of gapping the next day.The only problem with this approach is you have to be watching the screen,at those particular times.Good luck.....
|
bert123
53 posts
msg #71690
- Ignore bert123 |
2/24/2009 1:28:10 AM
Thanks Decipherlinda and Mike. I've been using the simple gap up filter to look at those patterns which precede the gap up and, most importantly, how often that pattern appears overall. Just as in horseracing, where 80% of winners ran in the last 14 days, if 80% of all horses ran in the last 14 days, that stat is useless. I was thinking of using the gap filter and then applying a condition such as; closed in bottom 25% of range, closed above close 1 day ago, closed below close 1 day ago,
top 25% of range, NR4, WR7, doji, reverse hammer, etc. But then you have to also know how often the pattern appeared overall and that's where the coding experts come in. I don't believe that there is a one-only pattern but I do believe some are better than others and those are what I'm looking for - thanks again.
|
decipherlinda
133 posts
msg #71734
- Ignore decipherlinda |
2/25/2009 6:36:17 AM
Seems like your latest post showed you're looking for something a little different from your first. It's okay but we're just trying to figure out exactly how to help.
At first it seemed like you wanted info on gaps in total as related to candles, not necessarily related to particular stocks. In general we can't do in SF but relative to specific stocks, yes. You'd have to download the specific stock results to Excel to perform generalizations about gaps.
Your last post referenced particular horses (stocks) and I think Mike's recommendations of studying gap filters already in these forums would help you. I believe there's a filter to find stocks that have gapped up most frequently (horses who've competed the most, most recently). You could add to that various criteria such counts for the candle types and technical criteria immediately preceeding the gap that you're interested in.
For example, a filter could tell you that AAPL gapped up 17 times in the last 100 days and in five of those 17 times a long white candle (5% or greater from open to close) preceeded the gap up. Or, in 3 of the 17 times, the preceeding close was above the upper bollinger band, etc.
SF filters are limited in that there's only so much you can do at one time. If you can identify more precisely what you want and if you can start with maybe one type of candle or one type of technical situation preceeding the gap, you'll probably find Chetron will be your best friend. He's generous when it comes to helping SF members develop filters.
Still I ask, are you sure this is the best use of your time? If there is a useful association, would it already have been discovered?
|