|
PureBytes Links
Trading Reference Links
|
Part I:
RealTraders: ADMIN: RealTraders' What Works? The Guide No Vendor Wants You
to Read, Part I
After reading this, please email your articles and ideas to me at:
WhatWorks@xxxxxxxxxxxxxxx
You email on this topic sent to any other address will be ignored.
The industry that has grown up around trading is one of the sleaziest in
the world. I am often embarassed to be a part of it. Recently, I decided to
compile a RT member authored guide that will serve as kind of a rating
system to evaluate products such as trading software, systems, books, and
seminars. You might ask, why do this?
Well, while I think a trader's should be on doing his own tinkering, one
should not start from scratch. There are a lot of good ideas that can be
molded and adapted to suit your own trading. Thus, even a lone maverick may
be in the market from time to time to purchase a book or indicator of some
sort. As a trader for the past 25 years (I started when I was in high
school), I do a lot of tinkering but I will also use whatever technique or
approach that works regardless of whether we developed or not. Nothing
matters except bottomline results. As you might expect, I have found very
few approaches that work or perform as claimed.
I'm sure that the majority of RealTraders know this frustration on a very
intimate level. No wonder we hear people asking, "What Works?" all the
time.
My proposal is that we all pool our cummulative knowledge and come up with
objective ways of evaluating all the tools that are out there right now.
Examples:
1. The Turning Point Trap. You have different types of programs which are
designed to forecast future turning points. Market geometry approaches such
as Gann and Fibonacci fall into this category, as do some neural nets. Have
you ever heard any vendor selling such an approach define quantitatively
what a turning point is? If you haven't, then you are allowing a floodgate
of loopholes which enable a vendor to define any little burp in the market
to be defined as a turning point and thus claim 90% accuracy!
2. The Optimization Trap. Frequently systems are advertised with impressive
track records. The statistics look impressive. But look a little closer and
you'll see that the results are hypothetical and apply only to a particular
period of time. It is easy to get any system to work based on historical
results. Even "walk-forward" back-testing can be deceptive if the
back-tested historical and walk-foward data does not cover a very long
period of time (sorry 5 years doesn't cut it!). In testing our own stuff,
we prefer validate a methodology by testing a data pool that extend over
scores of decades through many different types of markets. But this
guarantees robustness.
3. The Trap of Guru Dogmatism. Very frequently, we see people trying to
build personality cults around themselves. You've seen this happen on
RealTraders and I hate it even more that you do. You don't know how much
time I've wasted on that in the past week. I believe we must shun people
who make specific, absolute sounding statements about what the market will
do in near or far future. Why? Because, even if they are right a good
percentage of the time, they take away our ability to be flexible in
response to what the market is doing right now. A case in point is that I
heard there was a guy on the misc.invest.futures newsgroup who is always is
bragging about his market predictions. Last week he came right out and said
that the market absolutely would not drop below a specific level (he named
the level). The mini-crash blew right though his support point. How many
weak minded traders made costly errors because of this irresponsible
statement?
My purpose in doing this is not to position myself or any one person in the
group as a crusader for consumer rights. Rather, I am hoping that this
piece will be authored for the most part by articles submitted by
RealTraders members.
Continued in Part II
Eddie Kwong
|