PureBytes Links
Trading Reference Links
|
Tom,
Is this what you want:
colB-colA<(colC-colB)*0.99 and
colB-ColA<(colC-colB)*1.01 ?
Regards
Paul
----- Original Message -----
Sent: Tuesday, March 15, 2005 1:13
AM
Subject: [EquisMetaStock Group] Triple
Moving Average
I've been working on Dawn Balton Smith's triple
moving average, which is a five, fifteen, thirty simple moving average
based on the close. Now one characteristic of this set of averages is that
when the averages are equidistant from each other or if they are
clustered, this is supposed to be a strong indication for a trend
reversal. I'm trying to write a Metastock exploration to find when the
averages are both clustered or equidistant. So far I have set each
average in column A, B, and C respectively ie. Mov(close,5,s) in column A,
Mov(close,15,s) in column B and Mov (close,30,s) in column C. In the
filters column I have ColB-ColA=ColC-ColB. This works but the problem is
that the averages are calculated to four decimal places, which is too
exact and the exploration doesn't come out with enough results. I need
to adjust it to get say a 0.5% margin for error, which would give the
averages a couple of ticks leeway. Does anyone have any idea how I can
loosen up the criteria a bit to get more results out of the exploration.
Any tips would be greatly appreciated.
Regards, Tom
Yahoo! Groups Sponsor |
ADVERTISEMENT
| |
|
Yahoo! Groups Links
|
|