Re: Screwed up again...tried to analyze shorts and found potential longs....

["Jim Greening" <JimGinVA@xxxxxxxxxxxxx>]

Steve,
     Seems like a reasonable plan to me.  One of my Telescan searches does
a similar thing.  It looks for stocks with good fundamentals that have had
20 to 80% pullback and are now rising.  I also agree that system tests are
only good for certain groups of stocks.  That's why I use five different
ones and 
delete stocks that don't show at least a 100% gain on one of the tests
using 1000 days of data.

Jim
----------
From: Steven Buss <sbuss@xxxxxxxxxxx>
To: Metastock-list <metastock-list@xxxxxxxxxxxxx>; Jim Greening
<JimGinVA@xxxxxxxxxxxxx>
Subject: Screwed up again...tried to analyze shorts and found potential
longs....
Date: Tuesday, October 07, 1997 4:03 AM

 Jim Greening and the list,

Thanks for the CRUS gif file.

I'm new to MS 6.5 and the list.  I've had thoughts about posting from time
to time but feel like I don't have much value to share.

I very much appreciate your postings.  I ran a couple of the system tests
you posted a couple of weeks ago.  They seemed to work well for up-trending
stocks, not so well for stocks in a downtrend.  My sense is that it's very
difficult to build a set of symetrical tests that work for both.  (c.f.
Chande in the latest TAS&C on building symetrical tests).  When I look at
patterns I always seem to subconsciously look for longs.  It's irritating.

I noticed that there are a large number of equities up 200 - 500% for the
year.  I decided that it would be useful and maybe profitable to try to get
at what recent topping patterns and tests look like so I could apply them
to current data.  So, I wrote an exploration that was intended to pull out
equities according to the following criteria.

The stock is currently at more than 10 and, during the last 300 periods,
the stock gained more than 300% and then declined more than 40%.  The
extremely simple specific Exploration Filter is:

C > 10 AND 
C / HHV(HIGH, 300) < .6 AND
HHV(HIGH, 300) / LLV(LOW,300) > 3

The more experienced among you will notice my error immediately.  Mostly
what was returned was a list of stocks for which the recent Low is 33% or
less than the more distant high.  I was interested to see that many, not
all, of these stocks had what I understand to be a classic basing pattern
after a decline.  (CRUS is not on the list, but I think would have been if
the period tested for was a little greater than 300.)

So, I'm thinking that I will temporarily turn my attention to these new
potential longs and figure out which ones I like with other indicators.

Thoughts?

Steve Buss

sbuss@xxxxxxxxxxx