[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: what are fractals



PureBytes Links

Trading Reference Links

John Adair wrote:
>Would someone please tell me  what are fractals and why should I know about
>them.
>

John,
That is a great question, a question that has been the subject
of entire books.

A simple answer is that "fractal" describes something that looks
more-or-less the same independent of the magnification that you
are using to look at it.

For example, have you ever looked at a 5-minute chart, a 15-minute
chart, a 60-minute chart, daily chart, and weekly chart and
noticed similarity between the charts?  This is an example of a
"fractal" geometry in time.

Another often-cited example is a coastline... if you look at it from
2 inches away, you will see the same type of features you see when
you look at it from 5 feet away, or 100 feet away, or from overhead
in an airplane, or even in picture beamed back from the space shuttle.

As applied to the markets, the word fractal usually implies that
what a price does today depends on what has happened to the price
over the past.  For example, maybe you notice that for a certain
stock, you will see a pattern repeat itself, like UP, UP, UP, DOWN..
i.e. every 3 UPS is followed by a down (this is a very simplistic
example).

But, the idea that a market can be fractal in nature is very
controversial... there is something called the "Efficient
Market Hypothesis" which basically says that the current price
takes into account all known information, and it doesn't matter
what happened yesterday.

My favorite books on the fractals as it applies to the market
were written by Edgar Peters... there is a website at
http://oara.org/mpc/fma
that gives some background on this, but my practical advice
is to stick with "simpler" notions like support and 
resistance.  I've spent a lot of time looking into fractal 
models and I do think they can be very worthwhile, but I don't
know if the amount of time needed to get a good understanding
isn't better spent on developing less sophisticated models.
[the code for William's "fractals" looks fairly simple, I'd
be interested to know if anyone has had any success at all
with it...I've grown somewhat cynical about a lot of
these indicators].

Jeff