|
PureBytes Links
Trading Reference Links
|
I do not remember where this thread started, but some may remember it on this
List.
I think that the nickname came from Bob Brickey, who said that if they work,
it could be only by chance.
This is probably right if you use any kind of trading model available as a
standard in TradeStation or Metastock [ replace here the name of your favorite
crisp rule's editor], and in most of the case for the codes published in some
well know technical analysis magazines.
Building a robust trading system is not an easy task, and there is no magic in
any indicator or combination of indicators.
The core problem for traders is that they believe some indicators to be better
than others.
This is not true because an indicator is nothing different from a measurement
of some specific characteristics of the underlying data series (raw data).
The misleading point is in the way to use them, where most of us fail because
the rules given with the indicators are too simple to encompass the market
complexity.
I have recently been told (after neurofuzzy sparse wild advertising on this
list) by some contributors:
"Now, Pierre, enough of that. If you have proofs, please explain Right now".
This is what I have done during some free time, and I have built a web page to
show you not only why it can work or how it may work, but how to make it work.
I started with the only TWO well-known indicators (Wilder's ADX and DMIPlus
DMIMinus difference), then the only THREE (Wilder's ADX, DMIPlus, DMIMinus).
Only with this, I was able (in one hour) to build and test one trading system
(trained on less than 5000 bars of 10 min. SPZ6 SPh7 data) that generalizes
quite well on SP 98 10 min., Matif CAC40 1997, 1998 10 min futures and on
Daily commodities (100 years tested, also compared with the classic DMI
crossover).
I have spent several days to build a web page (huge size, many graphs,
comments) only to explain this.
Nothing was hidden in the way I worked to build the system. So, it could be
reproduced by anyone.
The URL is
<A HREF="http://www.sirtrade.com/fuzproof.htm";>
http://www.sirtrade.com/fuzproof.htm</A>
Please check it if you are interested.
I bet that you will be somewhat astonished that so stupid indicators were able
to generalize on more than 40,000 bars of unseen data.
Of course, the indicators were not optimized, because we cannot (we optimize
the trading rules only).
100% bull, you think? Not at all.
At the end of this month, we will release the Assistant for Expert Traders
Version 2.
A free evaluation will also be available, and with this one, you will be able
to produce similar results because we will allow testing on your own unseen
data (i.e. different than the Train/ Test data used to build it).
The kind of tool that could even change Neil Weintraub's mind...
Sincerely,
-Pierre Orphelin
|