ashraf999
33 posts
msg #46936
- Ignore ashraf999 |
9/14/2006 5:17:51 PM
I modefied it alittle bit. Does the following fit your criteria?
|
nikoschopen
2,824 posts
msg #46946
- Ignore nikoschopen |
9/14/2006 9:18:10 PM
ash,
First of all, you don't need to add "0 days ago" nor "offset 0" in ure filter. In the absence of any values (eg, 10 days ago, or date offset 10), SF takes 0 as its default anyway. Also the first line, "days(price 1 day ago crossed above center Keltner Band(7),9) equal 1", is a bit ambiguous. Literally it sez "As of 1 day ago, price 1 days ago crossed above todays's center Keltner Band(7)." Is that what you had in mind? If so, why?
You also need to take into consideration of the "double bottom" formation, which isn't very hard to code but it may be too limiting in scope. As a result, no stocks will be fetched. Here's a very primitive screener that I use in conjuction with this setup (yes, I have other filters, too):
As of today, 9/14, only one match (ASF) registered on the screen. But if you look at the chart of ASF, you know that it has yet to form a "W" or the "double bottom" formation. So you might want to keep ure eyes out on this stock for a few more days.
On a closing note, I would like to re-emphasize that this setup is more of a strategy than a system. As a daytrader, I apply this method on 5-, 15- and 30-minute charts. As long as you know what to look for, you need not bother chasing after myriads of different stocks. Since you can apply this method to any chart, you really need only one.
|
ashraf999
33 posts
msg #46948
- Ignore ashraf999 |
9/14/2006 11:08:44 PM
I am still getting confused between days(....) and count(....). Sometimes I think they are the same. After I changed it, I looked at the chart and found the "W" formatioan for "ESNR" as of 9/14. However, dma(28,-14) had already been shifted for the first crossover.
|
ashraf999
33 posts
msg #46949
- Ignore ashraf999 |
9/14/2006 11:11:21 PM
|
nikoschopen
2,824 posts
msg #46954
- Ignore nikoschopen |
9/15/2006 2:46:30 AM
ash,
Count() vs day() functions aren't as confusing as you might think. Count() will find the number of occurences within x number of days. Day(), on the other hand, will simply spit out how many days ago such occurences happened.
Going back to the above example, sometime the "double formation" (DB) is nearly indiscernible or nonexistent in certain stocks. Often, they will shoot straight up to the moon without making a DB. It's only fair to say that there's no way to know in advance which will make the DB and which will catapult without a pause. However, the risk you take with the ones that make the DB is a samall one. You know that if the trade turns against you, you can get out with a minimal loss.
It's also worth noting that the DMA(28,-14) turn upward in a timely manner. It should go north within days, if not immediately, once you embark on a trade. As long as it remains up, you may safely stay put until the center Keltner crosses above the DMA(28,-14), at wich point ure wise to close shop.
Since a picture is worth a thousad words, I'll see if I can find a good looking chart and upload it to a image hoster.
|
ashraf999
33 posts
msg #47005
- Ignore ashraf999 |
9/17/2006 2:58:23 AM
I am not sure if there is magical in this filter that is based on dma(28,-14) and keltner(7) but it seems to me the price keeps going up for 2-3 days at least.
|
nikoschopen
2,824 posts
msg #47064
- Ignore nikoschopen |
9/19/2006 4:15:13 PM
LOL. There is no magic in this setup but what more can I expect as a day trader? Ever since ure request, however, I've looked through many charts on a daily interval and found many stocks that rallied for weeks. Hence, I'm convinced that this strategy is not limited only to day traders.
|
nikoschopen
2,824 posts
msg #47065
- Ignore nikoschopen |
9/19/2006 4:48:01 PM
Having said that, I should point out that there are a few lines in ure filter that need some adjustment. First, you want either a flat or ascending rather than a descending DMA. It's also not necessary to see a close above the DMA, but you do want a white candle (ie, close above open) above the center Keltner. The 9th line that reads "low below lower keltner(7)", although not contradictory to my premise that close be above the center Keltner, is highly unlikely to occur. Hence you might want to omit it altogether. All in all, however, I'm glad to see that ure filter is taking shape.
For a visual illustration, refer to
http://img97.imageshack.us/my.php?image=patternss5.png
I haven't included a detailed annotation. I would rather have you look through the chart in detail and comment on what you think to be a discernible pattern pertaining to the strategy in question. If you want to reach in private, feel free to contact me at [my username]at<gmail.com>.
|
gocats
22 posts
msg #47170
- Ignore gocats |
9/26/2006 4:06:47 PM
Hey Nikoschopen
Here's what I come up with (My first filter)
I like the results that its giving me. Do you see anything wrong or anything that should be changed?
|
nikoschopen
2,824 posts
msg #47172
- Ignore nikoschopen |
9/26/2006 4:35:49 PM
gocats,
By glancing over the charts using the "date offset" function alone, the results looks pretty damn good. Could you explain what the rationale is for including the following lines?
set {cPerc2, upper keltner(7) * 1.5 }
set {cPerc, cPerc2 / 100}
set {cLow, upper keltner(7) - cPerc}
and ma(28) > cLow
Mathematically speaking, I could see that ure taking the difference between cPerc and the upper Keltner(7), but I don't understand where the MA(28) fits into the overall scheme of things. Perhaps you see things differently than this poor soul. :^)
|