|
PureBytes Links
Trading Reference Links
|
This post is an update to a previous post sent to the RealTraders
forum in April this year regarding PatternSmasher - a software-
based technical analysis package for stocks and commodities.
PatternSmasher (PS) uses proprietary technology to recognize
user-defined chart patterns. It's the 2000's approach to Edwards
and Magee style chart analysis. PS was released to the public
early this year by Ed Kasanjian of Kasanjian Research in Blue Jay
CA. The product has been advertised in Technical Analysis of
Stocks and Commodities along with Kasanjian's first analysis
package "Nature's Pulse"...designed to forecast turning points in
stocks and commodities.
The reason for this post is that I'd like to share my experience
using PS with others. I have not seen any reviews on PS
to date...although I admit that I have not looked too hard for
one. There have been many software packages I've purchased in my
life in which I wished someone had shared their experience with
the software BEFORE I purchased it. It is in this spirit that I
submit my thoughts and opinions. Secondarily, Ed Kasanjian
appears to be too busy updating his current software to work on
getting it widely reviewed. It should be noted that the product
has undergone SUBSTANTIAL changes since my April 1999 post
(below). For this reason...I will briefly summarize the current
PS product to bring things up-to-date regarding the latest 2.01
release of PS last week. To avoid any possible misunderstanding,
allow me to clearly state that I have absolutely NO connection
to, or interest in, Kasanjian Research...nor have I been asked to
submit posts on their behalf.
NOTE: Those that did not read my April post regarding PS may
prefer to read it first before reading this update. Kindly refer
to the RealTraders Forum post at the end of this update.
It appears that those having the most success with PS are willing
to devote time to developing and refining their own patterns
(called "PatternBots" or PB's). Each PB has the ability to locate
the specific pattern it was designed for. The early versions of
PS came with factory-designed PB's as design examples. The new
version of PS comes with no built-in PB's. This is a result of
early PS users believing that PB learning examples should turn PS
into a "push-button" money-making machine requiring no real
thought or effort to use "as-is" for profitable trading. More
than a few PS owners have been disappointed in this regard.
Any software bugs in the current version of PS appear to be minor
and easily fixable. The speed with which PS analyzes most complex
patterns has been substantially increased. Actual speed will vary
depending on the nature of the pattern. All basic functionality
one should expect from PS is now available - including extensive
group management of numerous issues, the ability to easily run an
analysis of a previously run analysis (looking for patterns
within patterns), basic printing functionality, and the ability
to run automated analysis on thousands of stocks/commodities
daily to locate rare and/or desirable time-price patterns.
With the latest release, there are far fewer limitations as to
the types of patterns that can be "programmed". A real expert in
PS will see an interesting pattern in a stock or commodity, draw
it out on paper, and then spend 30 minutes to several hours to
"program" PS to search for the desired pattern. The programming",
or PB design, of which I speak required no programming
experience. It should be noted that PB design with PS is not
intuitive. Contrary to a lot of "shrink-wrapped" software these
days...you MUST read the manual and/or take the free lessons
offered on PB design. Although PB development time is a direct
function of the complexity of the desired pattern...the
developer's pursuit of perfection...and the user's expertise in
using PS. Those unwilling to spend a couple of solid weeks
learning PB development will have difficulty using PS
successfully. It's a learning and iterative process to discover
what works and does not work with regard to market patterns...and
the PB's used to identify these patterns.
There remain several features that expert PS users still seek
from PS. Currently, data output is largely graphical with regard
to dates, prices, entry points, etc. Text printouts are available
for listing stocks in a particular group, or stocks that meet a
certain pattern or filter criteria. PS offers no built-in
backtesting facility or clear trade exit system. Put simply, PS
is a tool...and not a system. I suspect many users that are
unable to develop their own separate trade exit strategies (to
accompany PB triggered trade entries) will receive little benefit
from PS.
Another much requested feature that was not in the current
release is data compression. In other words, the ability to take
daily data and compress it to do analysis on weekly or monthly
data bars. PS only works with end-of-day data at this time. PS
users continue to wait for the real-time version of PS which
should, in theory, handle any data compression scale from one
minute to a month or more. Daytraders may gain benefit from the
longer-term forecasts provided by PS. Unless one is quite
computer savvy and can trick PS into believing that 5-minute data
is really daily end-of-day data...I do not see any way to use PS
in real-time daytrading in its current form. It is not yet clear
how fast PS will perform as a real-time trading engine whenever
real-time data feeds are supported in a future version.
In summary, PS may well be amongst the most powerful weapons in a
technician's arsenal. It is not, however, for those that do not
understand the basics of technical analysis...or...those that
cannot devote time to incorporating PS results into an overall
trading strategy. I suspect that at least half of PS users simply
use it to prequalify stocks for further analysis by a more
comprehensive system. I wouldn't want to give it up for any
amount of money...but I also would not want to have it alone with
no understanding of trade entry/exit strategies, money
management, and fundamental technical analysis principles, etc.
PS is definitely NOT the "holy grail"...but if anyone ever
discovers it...PatternSmasher may well be one of the key
elements. The spirit of Edwards and Magee is still alive...in
PatternSmasher.
Andre Franklin
WickedTrader, LLC.
"Where outsiders trade like insiders"
http://www.wickedtrader.com
Email: wickedtrader@xxxxxxxxxxxx
________________________________________________________________
Posted to RealTraders Forum of April 9, 1999
I have listened to many comments on PS both here and on
PatternSmasher's private discussion group. I posted a message
several weeks ago to the PS group, but thought I'd also post to
RT as a discussion of PatternSmasher has recently surfaced. I am
having a very positive experience with the product...but I also
understand where it may not meet the needs of many.
I purchased PatternSmasher when it was first made available in
January this year...and I would not sell it back to Kasanjian
Research for several times what I paid them for it!!! I admit
that I am very biased in my opinions as I favor pattern-based
trading...and that is the real strength of PS. Personally, I like
systems that do not require optimization....and pattern-based
trading, in the spirit of Edwards and Magee, allows me to trade
with consistent results using patterns that were often considered
too subjective to trade consistently. PatternSmasher simply
allows me to take subjectivity out of my pattern
assessment...thus ensuring consistency in mechanical trading.
I do not want to allude that pattern trading is the only trading
strategy that does not require optimization. I will also admit
that patterns can be designed that are optimized for specific
issues and/or time frames....so the dangers of optimization still
apply. Still, I find that certain patterns seem to have general
validity across various issues and time frames....and such
patterns are "self optimizing" in a sense. A big triangle is no
different than a small triangle. The smaller triangle may have
developed as a result of a smaller time frame...and may not have
as much of a price move as the larger triangle....all things
being equal of course. Still, both triangles have the same
characteristics with regard to price moves that can be expected
at the completion of the triangle pattern. The market itself
determines how a triangle will develop...its size (in time and
price range), lead-in characteristics, etc. The resulting
triangle reflects the environment that caused its formation...and
thus it is "self-optimized".
One can calculate the probability of a certain price move in a
specific direction on the completion of a specific predefined
pattern. All technical analysis is based on probability
anyway...its just that pattern approaches look toward specific
shapes versus indicators, averages, etc.
For years, I was reluctant to base my trading too heavily on
patterns as I found them very difficult (but not impossible) to
trade in a 100% mechanical fashion, especially when I wanted to
pre-qualify many issues daily...and then pick the "best-of-the-
best" to trade. I am the kind of trader that needs a mechanical
system as I have never possessed the discipline to properly trade
systems that require trader discretion. I find a conflict of
interest for myself when it comes to trading my own money and
leaving room for emotions to affect decisions that should be
objective and business-based. I take my hat off to traders that
have won this battle in the absence of a relatively mechanical
trading system. For me...I need the discipline...and mechanical
systems give me that discipline.
Because PS allows absolute and objective quantification of
patterns, it has opened up a whole new world for me to develop
100% completely mechanical systems using pattern analysis as the
core decision-making engine. With PS, I found that I could do
research into classic patterns (i.e. - head and shoulders) and
understand more about how well the pattern works over a weekend
than Edward's and Magee may have learned in a lifetime! This is
not an exaggeration! The power of the computer to search for
patterns through thousands of stocks/commodities over many many
years, and with perfect objectivity, is almost miraculous. I
doubt that Edward's and Magee ever dreamt that there would be
such a day as this!
I'm not trying to say that PatternSmasher is for everyone. It is
not. The current backtesting facilities in PS are graphical only
(no data file output or TradeStation-like reports as yet), so
many will have a problem based on this fact alone. In hindsight,
however, I found that the lack of historical statistics forced me
to look at each and every pattern carefully. Inadvertently, I
learned about what worked and didn't work in patterns...and was
able to make certain adjustments and refinements to some of the
classic E & M patterns that have shown results I consider to be
very impressive.
There are a few things that some would consider to be "negatives"
to PS. The program does not force or recommend a specific entry
or exit point, but "automatic" trendline entries are provided on
break of a trendline drawn between pivots selected by the user.
Again, all of this is graphical at present. There is no facility
for trailing stops. Although I would advise cautious use of
backtesting statistics if PS were to offer this, the current
program needs to develop in this area (and KR has promised to
work on this). The bottom line, however, is that those purchasing
PS either need to be serious about pattern-based trading and/or
research, OR, they need to have reasonable intelligence and a
willingness to learn about patterns in markets. One should not
expect to open the box, press a button, and money fall out the
other side. PS is a tool...with more emphasis as a research tool
than trading system in the current version. Still, it can be
effectively used as a trading tool with a bit of effort and
reasonable entry/exit/money-management principles. I have done
this...so I will not listen to those that say it cannot be done.
PS is the kind of program that will either turn you off, or,
you'll find it the greatest thing since sliced bread. I admit
that not all trading strategies are for everyone....even if they
are very profitable. One must believe in one's system, else one
may be tempted to trade it differently than how it was designed
to be traded. For the same reasons that trading based on
fundamental analysis does not agree with me personally...I must
also allow others the same freedom to choose what works for
themselves. I'm convinced that two persons trading the same
identical pattern in PS will likely have different results.
Because one does not have good results should NOT invalidate the
results of others having better success. How many students did
Gann teach?....And how many have shared the same success as Gann?
To respond to a few questions on PS...
PS does, in fact, have features found on no other software.
Specifically, this is the ability to do serious research on a
variety of patterns for the first time in history. I'm not sure
that one can actually program ANY pattern that the "experts" have
taken years to study and learn, but I think a fair statement
would be that over 90% of such patterns can be programmed
completely, with perfect objectivity, assuming reasonable study
of the manual and the special tutorial lessons that come free
with the program.
Is PS worth the price? It depends on whether pattern-based
trading agrees with you or not. I would not trade a system that
was tested to make 100% profit per week if it did not agree with
me on the inside. I would advise anyone that is willing to put in
a reasonable effort and who is not completely satisfied with
their current system/strategy to give PS a fair try...and return
it if it does not work for you. I do not think most will be
disappointed, but for many, it will be the only way to know for
sure if such a tool fits both their style and profit objectives.
As I mentioned, I am having very positive results with PS. Some
might argue that it is my entry/exit and money management
strategy that is responsible for this...and I'd probably agree
that this is a big part of it. Still, in my twenty plus years as
a student and trader of the markets, I've never been able to put
together a system as profitable as what I now have using PS at
the core of the system.
Is PS a valid and viable approach? To this I'd have to say that
it is probably no more valid than Edwards and Magee and the
trading masters of this century that employed pattern-based
approaches to trading. Rather than ask, "Is PS profitable?", it
would be better to ask, "Are pattern-based trading approaches
profitable?" Regardless of the response to this question, there
can be no doubt that PS is the first product in history that is
uniquely prepared to address such questions. As "primitive" as
some may consider the PS graphics-based backtesting
facility...there still is nothing publicly available that can
provide as rich and deep an understanding into the behavior of an
innumerable variety of patterns as PS. In this regard, PS is in a
class by itself...and there is no fair comparison possible with
any other software. Yeah, I suppose it's possible to program
TradeStation to recognize a plethora of patterns, but to make it
do what PS does would require TS code with the complexity of
UNIX, WindowsNT, and Oracle databases combined!!! (or a
PatternSmasher TS module!). Put simply, PatternSmasher's pattern-
recognition capabilities are not possible in any software
available today. One must develop software, from the start, with
a foundation that understands the basic elements of patterns at
its very core. Such is the paradigm shift in technical analysis
software as represented by PatternSmasher.
All PS does is objectively find patterns...and can scan hundreds
of stocks/commodities daily without having to manually search
each chart. This feature alone can benefit the most seasoned
pattern trader as one can scan for the "best-of-the-best"
patterns amongst a thousand...and then cherry-pick amongst them
for the best trading candidates. Some will find PS to be the key
to consistent profits. Others will, no doubt, discover that they
cannot put together a whole system focused on PS, but will find
it useful as a filter in other trading strategies. And of course
there will always be those that cannot make a profit with PS to
save their life...even using the same patterns as others that are
having success. The differing success may be due to discipline,
money management, or even one having an account that could not
withstand a certain drawdown while the other trader's account was
well-funded.
I am not paid by KR, nor have I received or been promised any
special benefits to share a "glowing" review of PS with this
group. I have never met Ed Kasanjian or any of the KR staff. I'm
only sharing this in the spirit of the rich sharing I've received
on this discussion group and elsewhere over the years. There are
many software programs I wish I had read about BEFORE I mailed a
check!!! I'm hopeful that this will encourage others to openly
share their experiences with software programs - good or bad.
In summary, I simply want others to know that this user is having
a very very positive experience with PatternSmasher...but I
quickly acknowledge that the program may not be for everyone.
Andre Franklin
WickedTrader, LLC.
"Where outsiders trade like insiders"
http://www.wickedtrader.com
Email: wickedtrader@xxxxxxxxxxxx
|