Rodney,
I’ve
seen from your posts in the ICE users group why you are asking this question
and I think you might be making things harder for yourself than you need to.
The first thing though, is that you should take a look at what Superfrag has
posted recently regarding developing systems for up and down trends. Won’t
repeat it here, just to say that everything he says makes perfect sense –
both in theory and, most importantly, in practice. It’s the best ‘advice’
post that I have seen posted on this or any other group.
Secondly, if
you are looking to define historical trends mathematically so that you can then
pick periods where there is equal market bias and then test your system on it,
you’re putting the cart before the horse somewhat. Again, read what
superfrag has said – have seperate up-trend buy and down-trend sell
systems and put your energies into determining how and when to trust those
systems and when to ignore the signals. In a word…”filters”. For
back testing, the best tools you can use to determine a suitable period are
your own eyes and, if you really need some confirmation of what you’re seeing,
a couple of moving averages. There really is no need to get more scientific or
complicated than that. Your own brain contains the best pattern recognition software
there is…why try and replicate its function for something so simple as
picking out historical trends, or lack thereof, for back-testing? Get that
highlighter out man!!
From:
platt_rodney [mailto:platt_rodney@xxxxxxxxxxxx]
Sent: Sunday, January 23, 2005
12:50 PM
To: equismetastock@xxxxxxxxxxxxxxx
Subject: [EquisMetaStock Group]
Hindsight trending
Is there a way to identify the trend of historical
data? I know it
sound strange, why would you want to know the
trend of historical
data? Because I am trying to identify a period of
data with a good
balance of up trend, down trend and sideways
moving market data.
So if any one has a way(s) to identify the market
trend in
hindsight, beside getting the old highlighter out
and drawing on the
monitor it would be much appreciated.
Regards
Rodney
Yahoo! Groups Links