TheRumpledOne
6,407 posts
msg #69068
- Ignore TheRumpledOne |
11/6/2008 6:05:19 PM
This is a ONE TRADE PER STOCK PER WEEK SYSTEM.
|
ddavid
10 posts
msg #69157
- Ignore ddavid |
11/10/2008 5:42:28 PM
It seems to me that profits can be maximized (especially after a generally down day like today) by looking for stocks that were high on the weekend slow cooker list and that have not gone up 1% from the open today and closed considerably down. For example, every week out of the past 52 weeks RAS has closed on Friday up at least one percent from the open on Monday. Today, RAS opened at 3.15 (and it only needs to get to 3.18 to make 1%) but it never went over 3.15 and it closed at 2.43 (22.86% down). This means that chances are, RAS will still end up getting to over 1% above where it opened this morning. So if you bought it at 2.43 and held it to 3.18 (1% over the open today), that would be a massive gain! Even if you wanted to buy it at 2.43 and be "conservative" and assume it will at least get back to the open this morning, you would make 22.86%... has anyone thought about this idea? And is there any filter for returning such stocks so that we could see where the largest potential gains are after Monday? I am fairly new, so I don't want to "reinvent the wheel" so to speak, but if others think it would be useful, and nothing like it exists yet, I will certainly work on creating a filter for us.
|
maxreturn
745 posts
msg #69158
- Ignore maxreturn |
11/10/2008 5:58:33 PM
ddavid, seems to me that the BEST filter for utilizing TRO's fine filters is the broad market direction. When in a long term uptrend focus primarily on trading stocks from the long side. Hence, you use the long version of "Slow Cookers". Vice versa for a long term downtrend.
|
ddavid
10 posts
msg #69160
- Ignore ddavid |
11/10/2008 7:20:14 PM
I am not looking to determine if I should be long or short in this market... I am looking to build upon these filters to make them helpful for other days (after Monday)... so that profits could be maximized further for those stocks that fit my specifications (having not breached the up 1% from the open point) and for which you can now buy at an even cheaper price than at the open on Monday...
|
ddavid
10 posts
msg #69274
- Ignore ddavid |
11/18/2008 9:07:51 AM
I have been reading the StockFetcher tutorial for filters, and it has confirmed what someone from the stockfetcher support group has told me, which is that the count function does NOT support week counts, just day counts...
Therefore statements like this:
count(Long_Profit > .04 , 52)
are actually counting the number of times that the Long_Profit has been over 4% for the past 52 DAYS not weeks... so doesn't this ruin the theory behind this filter?
When I asked the support team how to create a filter such as what you are trying to do in this forum, the response was "Unfortunately, the feature you are looking for is not possible at this time. The "count()" function currently does not support weekly measures, it is based on days only at this time." ... So it seems that what you are trying to do with slow cookers cannot be done in stockfetcher at this time.
|
chetron
2,817 posts
msg #69275
- Ignore chetron |
11/18/2008 9:20:46 AM
SO EVEN THOUGH LONG_PROFIT IS THE RESULT OF WEEKLY DATA, THE SAMPLE OF 52 IS FOR DAYS AND NOT RESULTS?
|
ddavid
10 posts
msg #69286
- Ignore ddavid |
11/19/2008 8:51:54 AM
TRO - can you explain how your usage of the count function is counting the "slow cooker" pattern over the past 52 weeks since the count function only works for number of days as shown below?
The parameters to the count function, as described in the user guide is this:
Parameters:
Indicator (the indicator to apply the count to)
Length – (number of days)
Usage:
Count(Indicator, Length)
Therefore, this little snippit of code from the first page of the forum, seems to be finding when the (weekly high - weekly open) / weekly open is over 1% over the past 52 days, not over the past 52 weeks...
set{whiop, weekly high - weekly open}
set{Long_Profit, whiop / weekly open }
set{onepct, count(Long_Profit > .01 , 52)
TRO, If you could shed some light on this it would be great. TIA!!!
|
chetron
2,817 posts
msg #69287
- Ignore chetron
modified |
11/19/2008 9:15:54 AM
SO IT IS ONLY LOOKING AT THE LAST 10.4 WEEKS. 52 DAYS / 5 DAYS A WEEK EQUAL 10.4 WEEKS?
|
ddavid
10 posts
msg #69291
- Ignore ddavid |
11/19/2008 10:25:08 AM
no chetron... using the count function in this way makes no sense at all (according to the support team at stockfetcher)... hence why I am hoping TRO will explain his logic behind his code...
|
TheRumpledOne
6,407 posts
msg #69305
- Ignore TheRumpledOne |
11/20/2008 12:45:05 AM
Perhaps I am mistaken.
Hey StockFetcher, what's the correct answer?
|