PureBytes Links
Trading Reference Links
|
Ross -
Actually, Kaufman's book is very heavy on statistics. I used to do a
good deal of work on that sort of stuff too and had created trend
indicators based on various derivatives of regressions, etc. Also,
Advanced Get, which is an Elliott program is statistically based as are
many other technical systems. For that matter, aren't nerual nets by
definition probalistically based?
Steve Poser
-
Steven W. Poser, President
Poser Global Market Strategies Inc.
http://www.poserglobal.com
Ross Kovacs wrote:
>
> A recent post by Brent included this comment:
> "I got a copy of Perry Kaufman's Trading Systems and methods from my
> library. I was surprised how much this book covers. It makes me realize that
> almost every conceivable angle has been covered in the pursuit of trading."
>
> I must aver. I have a friend who manages over $100 m in a fund that has
> clients including Coca Cola's pension plan. My friend has a Ph.D. in math,
> specializing in statistics. He makes his investment decisions strictly on
> statistical analysis. I can never get him to describe specifics other than
> trades he is making. But some conversations have been borderline in
> allowing me to see inside his "box".
>
> His statistical methods allow him to switch between what we might describe
> as "trend following" to "fading the trend" strictly based on his
> probabilistic calculations. He would describe his fading trades as "waiting
> for the price to be in the tail of a probability distribution."
>
> When I discuss standard technical analysis indicators to him such as RSI or
> Bollinger Bands that might give the same buy/sell signal, he listens but not
> with great interest. He claims to have negligible knowledge of technical
> analysis and wants to remain that way. Perhaps a personal choice, perhaps a
> business choice since many pension fund managers view technical analysis as
> voodoo.
>
> Although I haven't read Perry Kaufman's book, I'd be surprised to see a
> purely statistical method in the book. I haven't seen a purely probability
> and statistical based method described anywhere, and my study includes an
> MBA in Finance as well as many years interest and use of technical analysis.
> One reason for this post is to ask any RTs have they seen or read of any
> methods based purely on probability and statistics? (I don't include Curtis
> Arnold's PPS in this category since that does rely on pattern recognition, a
> standard technical analysis category).
>
> A second reason is to try to open further debate. Anyone else seen or heard
> of methods "outside of the box" of methods that technicians or
> fundamentalists are familiar with?
>
> P.S. No flame intended. Brent used excellent etiquette by saying "almost
> every conceivable angle." My belief is that there are many new angles to
> be discovered. Unfortunately, many traders are secretive (including my fund
> manager friend) and don't reveal their "new" approaches.
>
> Ross Kovacs
>
> rossrk@xxxxxxxxxxxxxx
-
|