|
PureBytes Links
Trading Reference Links
|
There is a wider audience for these posts than just the people who
actively contribute.
My purpose was not to encourage a discussion on the subject, as it's not
completely appropriate for this forum. My intent was to help guide and
warn traders that aren't rocket scientists with a solid grounding in
statistics (virtually all of them).
I'll post two quotes from a guy much smarter than me on the subject, and
end my participation on this topic.
S.
"Research that starts with concepts is much more likely to avoid
curve-fitting than blind number-crunching."
"The philosophy of science teaches that all observation is theory-laden;
there is simply no way to analyze data in a theoretically neutral
manner. In fitting to historical data, theoretically unsound procedures
can lead to radically invalid conclusions. This is probably why
back-testing has developed such a bad reputation."
Richard Eckhardt
--- In amibroker@xxxxxxxxxxxxxxx, "wavemechanic" <fimdot@xxx> wrote:
>
> "Theorize, and then test the theory for validity."
>
> If you want to play in the trading world's "scientific method
sandbox," propose a falsifiable hypothesis and let all chew on it.
>
> Bill
>
> ----- Original Message -----
> From: "sebastiandanconia" sebastiandanconia@xxx
> To: amibroker@xxxxxxxxxxxxxxx
> Sent: Saturday, October 21, 2006 11:10 AM
> Subject: [amibroker] Re: Scan for Bulkowski's chart patterns
>
>
> > "...I suggest it is not due to limitations of our equipment, but
simply
> > a system that is no good as it is explained..."
> >
> > That's precisely the conclusion I've come to, and it's all Tomasz's
> > fault for building AB and making such testing possible.:)
> >
> > If I can strain your chess analogy a little bit, the edge that
> > grandmasters have is that, of all the potential moves available to
them
> > they can quickly recognize the critical ones and focus on them,
> > excluding all the others as unimportant "noise."
> >
> > As traders, that's what I think we should aspire to. A stock
doesn't go
> > up because a 5-day EMA crosses above a 20-day EMA or because the
> > directional movement index is positive and the ADX is greater than
30.
> > Those indicators coincide with the stock's uptrend but aren't the
cause.
> > Neither is the cause to be found in linear regression, Gann angles,
> > Fibonnaci numbers, etc. What is the cause, then? Theorize, and
then
> > test the theory for validity. That's when we can start to learn
> > something. Otherwise, all we're doing is testing random data for
> > patterns that are only meaningless coincidences, like the Superbowl
> > Theory.
> >
> > And that's enough sermonizing from me.:)
> >
> >
> > Luck,
> >
> > Sebastian
> > --- In amibroker@xxxxxxxxxxxxxxx, "M Webb" the_bear_98@ wrote:
> >>
> >>
> >> In the old days, programming a computer to play chess and capture
the
> > mind of a Grandmaster, with the instant "pattern recognition" they
> > possessed was very difficult. It took many years before the
combination
> > of computing power and programming finally got rolling. Now we have
> > programs that will beat you like a borrowed mule. It is easy to
grade
> > these programs as they have ratings that prove how good they are.
During
> > this same time frame, we now have the computers and programming to
> > really back-test and prove a trading theory might have worked in the
> > past. We can go buy a book, program it, and discover it doesn't
produce
> > a reward/risk worth trying at all. Now in the old days, the
Grandmaster
> > would say "I only think one move ahead, but it is always the best
move."
> > We found a way to mostly brute force find the same move, but when we
try
> > to program one of the successful traders systems, the back-test does
not
> > pan out. I suggest it is not due to limitations of our equipment,
but
> > simply a system that is no good as it is explained.
> >>
> >> Although it might be theoretically possible to combine two losing
> > systems with a magic ranking process that could produce an overall
> > winner, I feel safe in saying no one can come up with an example
they
> > would share.
> >>
> >> Monty
> >>
> >> ----- Original Message -----
> >> From: allansn@
> >> To: amibroker@xxxxxxxxxxxxxxx
> >> Sent: Friday, October 20, 2006 12:44 PM
> >> Subject: Re: [amibroker] Re: Scan for Bulkowski's chart patterns
> >>
> >>
> >>
> >> Just out of curiousity,is it possible to have two non correlated
> > trading techniques that show negative backtest results
individually,but
> > when combined produce above average returns...
> >>
> >> I am over in Elite trader observing a candlestick debate,and
trying
> > to understand how "savvy" traders who fully acknowledge that
> > candlepatterns fail miserably in backtests insist that they are a
great
> > confirming indicator..
> >>
> >> Is there a method to this madness??
> >>
> >> Allan
> >>
> >>
> >>
> >> ----- Original Message -----
> >> From: sebastiandanconia
> >> Date: Friday, October 20, 2006 10:33 am
> >> Subject: [amibroker] Re: Scan for Bulkowski's chart patterns
> >> To: amibroker@xxxxxxxxxxxxxxx
> >>
> >> > I've read about this kind of behavior before, and it's a
problem
> > for
> >> > anyone who is trying to emulate the methods of successful (or
> >> > allegedlysuccessful) traders. What is happening is that these
> >> > traders aren't
> >> > fully conscious of everything in their decision-making process.
> > There
> >> > are "soft" factors that are key to their success that AREN'T in
> > their
> >> > "written down" or "official" methods. And if every factor isn't
> >> > included the method can't be accurately tested for validity.
> >> >
> >> > JMO, but I suspect that's one of the reasons it's so difficult
for
> >> > traders to duplicate the success of even legitimate,
well-meaning
> >> > system-sellers or "gurus." Maybe they really know something
> >> > unique, or
> >> > maybe they just had a run of luck when their "fuzzy" thinking
> >> > paid off
> >> > but that particular set of conditions won't ever be repeated.
> >> >
> >> >
> >> > Luck,
> >> >
> >> > Sebastian
> >> > --- In amibroker@xxxxxxxxxxxxxxx, allansn@ wrote:
> >> > >
> >> > > Hi Jim,
> >> > >
> >> > > What I find interesting about various pattern recognitions
> >> > such as Edwards and Magee type formations and candlesticks is
> >> > when you
> >> > backtest them in isolation they simply show no predictive
> >> > power.I am
> >> > predominantly refferring to Candlesticks as I have done
extensive
> >> > testing in that area.
> >> > >
> >> > > Yet,I have friends who strictly trade off them with western
> > price
> >> > patterns and trendlines and return well over 50% annually.
> >> > >
> >> > > Others I know use them as a filter to their systematic
> > trading,and
> >> > they know full well that Candlesticks do not backtest well.But
> >> > they wont
> >> > trade without them
> >> > >
> >> > > Interesting how the mind works
> >> > >
> >> > > Allan
> >> > >
> >> > >
> >> > >
> >> > >
> >> > >
> >> > >
> >> > >
> >> > >
> >> > >
> >> > > ----- Original Message -----
> >> > > From: jwmc107
> >> > > Date: Thursday, October 19, 2006 7:48 pm
> >> > > Subject: [amibroker] Re: Scan for Bulkowski's chart patterns
> >> > > To: amibroker@xxxxxxxxxxxxxxx
> >> > >
> >> > > >
> >> > > > Hi Joe, Brian, Alan and others,
> >> > > >
> >> > > > Here are Tom's trading results:
> >> > > > Year 2000 2001 2002 2003 2004 2005
> >> > > > Tomas 14.80% 18.70% 9.00% 13.50% 29.10% 15.40%
> >> > > > S&P500 -10.10% -13.00% -23.40% 28.69% 10.88% 4.91%
> >> > > > Nasdaq -39.30% -21.10% -31.50% 50.00% 8.60% 1.40%
> >> > > > DJIA -6.20% -7.10% -16.80% 28.28% 5.31% 1.72%
> >> > > >
> >> > > > The results speak that the system works. Hope that someone
can
> >> > > > make
> >> > > > the code available, as I dont have any programming skill
> > myself.
> >> > > >
> >> > > > There are WL codes available, we can just translate it to
AFL.
> >> > > >
> >> > > > Regards,
> >> > > > Jim
> >> > > >
> >> > > > --- In amibroker@xxxxxxxxxxxxxxx, "Joe Landry" wrote:
> >> > > > >
> >> > > > > Allan
> >> > > > > I had to think back a bit. Well I'm schiz'ed, a
> >> > > > fundamentalist
> >> > > > and a technician. Making money on fundamentals right
> >> > > > now(Validea)
> >> > > > but I know what
> >> > > > > will preserve it is technical analysis.(or risk
management?)
> >> > > > >
> >> > > > > What might be an interesting side note, we had Nelson
> > Freeburg
> >> > > > of
> >> > > > Formula Research give a talk to us last Sat. and
> >> > > > > he related as to how his first investment book while he
> >> > was at
> >> > > > Columbia was Martin Pring's early TA book. This led to his
> >> > > > development of a first "system" called Primas (Pring Moving
> >> > > > Average
> >> > > > System). I've got a cam video of his talk on our server,
let
> >> > > > me
> >> > > > know if you'd like to down load it.
> >> > > > >
> >> > > > > Best regards
> >> > > > > JOE
> >> > > > > ----- Original Message -----
> >> > > > > From: allansn@
> >> > > > > To: amibroker@xxxxxxxxxxxxxxx
> >> > > > > Sent: Thursday, October 19, 2006 7:53 AM
> >> > > > > Subject: Re: [amibroker] Re: Scan for Bulkowski's chart
> > patterns
> >> > > > >
> >> > > > >
> >> > > > > Joe,
> >> > > > > I am shocked...I never would have taken you for a
> >> > > > closet "Edwards and Magee" man :)
> >> > > > >
> >> > > > > Allan
> >> > > > >
> >> > > > > ----- Original Message -----
> >> > > > > From: Joe Landry
> >> > > > > Date: Thursday, October 19, 2006 8:15 am
> >> > > > > Subject: Re: [amibroker] Re: Scan for Bulkowski's chart
> > patterns
> >> > > > > To: amibroker@xxxxxxxxxxxxxxx
> >> > > > >
> >> > > > > > I just dusted it off Dimitris' program again and was
> >> > > > stepping
> >> > > > > > through the
> >> > > > > > selections.
> >> > > > > > If you're not familiar with it you might want to check
out
> >> > > > > > Anthony
> >> > > > > > Faragasso's charting script
> >> > > > > > as labelled below and in the program library. It's a
good
> >> > > > > > companion for this
> >> > > > > > exploration.
> >> > > > > >
> >> > > > > > I tried my hand at on of his most successful patterns
> >> > > > called
> >> > > > > > "pipes" but was
> >> > > > > > not happy with the results.
> >> > > > > >
> >> > > > > > BTW Bulkowski has a newer paperback called Getting
Started
> >> > > > in
> >> > > > > > Chart Patterns
> >> > > > > > pub within
> >> > > > > > the last year.
> >> > > > > >
> >> > > > > > As far as Thomas is concerned, I would be a potential
> >> > > > customer
> >> > > > > > for a set of
> >> > > > > > Bulkowski pattern recognition
> >> > > > > > routines if such were to be produced and put up for
sale.
> >> > > > > >
> >> > > > > > Best regards
> >> > > > > > JOE
> >> > > > > > SECTION_BEGIN("Pattern Recog. Tony Farragasso");
> >> > > > > >
> >> > > > > > // Name and Purpose ----------------- Faragasso Pattern
> >> > > > > > Recognition -----------------
> >> > > > > >
> >> > > > > > // Date 11/23/03
> >> > > > > >
> >> > > > > > // Source Anthony Faragasso
> >> > > > > >
> >> > > > > > // Type Indicator, Exploration, AB PT
> >> > > > > >
> >> > > > > > // Patterns Wedge, Broad, Upchannel, Bearish.
> >> > > > > >
> >> > > > > >
> >> > > > > > ----- Original Message -----
> >> > > > > > From: "brian.z123"
> >> > > > > > To:
> >> > > > > > Sent: Thursday, October 19, 2006 12:50 AM
> >> > > > > > Subject: [amibroker] Re: Scan for Bulkowski's chart
> > patterns
> >> > > > > >
> >> > > > > >
> >> > > > > > > Hello Jim,
> >> > > > > > >
> >> > > > > > > Do you mean the chart patterns from Bulkowskis
> >> > > > encyclopedia
> >> > > > > > (over 60
> >> > > > > > > patterns)?
> >> > > > > > >
> >> > > > > > > If you want to code them in AFL some have been done
by
> >> > > > Dmitiris
> >> > > > > > > Tsokaki! s (Hello Dimitris, I hope the fish are
jumping
> >> > > > into
> >> > > > your
> >> > > > > > > boat)....from the AFL library.
> >> > > > > > >
> >> > > > > > > Formula name: Pattern Recognition Exploration
> >> > > > > > > Author/Uploader: Dimitris Tsokakis - tsokakis [at]
> > oneway.gr
> >> > > > > > > Date/Time added: 2001-09-03 14:03:35
> >> > > > > > > Origin:
> >> > > > > > > Keywords:
> >> > > > > > > Level: medium
> >> > > > > > > Flags: exploration
> >> > > > > > >
> >> > > > > > > DISCLAIMER: Most formulas present in AFL on-line
> >> > library are
> >> > > > > > > submitted by the users and are provided here on an
"as
> >> > > > is"
> >> > > > and "as
> >> > > > > > > available" basis. AmiBroker.com makes no
> >> > representations or
> >> > > > > > > warranties of any kind to the contents or the
operation
> >> > > > of
> >> > > > material
> >> > > > > > > presented here. We do not maintain nor provide
technical
> >> > > > > > support for
> >> > > > > > > 3rd party formulas.
> >> > > > > > >
> >> > > > > > > Download formula file | Delete formula
> >> > > > > > >
> >> > > > > > > Description:
> >> > > > > > > Paste the code in Automatic Analysis and Explore All
> >> > > > Stocks
> >> > > > > > for the
> >> > > > > > > last date of your data.
> >> > > > > > > In Results click twice a label , for example "a TR"
and
> >> > > > you
> >> > > > will
> >> > > > > > > collect all ascending Triangles marked with an "1".
> >> > > > > > > This Exploration recognizes the following patterns:
> >> > > > > > > +Ascending Wedges "aW"
> >> > > > > > > +Descending Wedges "dW"
> >> > > > > > > + ascending Triangles "a TR"
> >> > > > > > > +descending Triangles "d TR"
> >> > > > > > > +ascending channel "a ll"
> >> > > > > > > +descending channel "d ll"
> >> > > > > > > +almost horizontal channel "ll"
> >> > > > > > > +Converging Support and Resistance
> >> > > > > > > +Diverging Support and Resistance
> >> > > > > > > +N/A, ie examples not belonging to any above
category.
> >> > > > > > > You may decrease sensitivity, changing
> >> > > > > > > per=3;
> >> > > > > > > with another percentage higher than 3.
> >> > > > > > > For Indicator Builder paste the code in Comments
> >> > > > > > >
> >> > > > > > >
> >> > > > > > > It would be an interesting project to code all of the
> >> > > > major
> >> > > > patterns
> >> > > > > > > (if it hasn'! t been d one already).
> >> > > > > > >
> >> > > > > > > There is also a formula there for identifying candle
> >> > > > patterns
> >> > > > > > by Top
> >> > > > > > > Gun Herman van den Bergen....
> >> > > > > > >
> >> > > > > > > Converting candles to bar chart patterns would be
> > another
> >> > > > > > > interesting project.
> >> > > > > > > I'll do both projects if I run out of things to do.
> >> > > > > > >
> >> > > > > > > I haven't tried either of the formulas....I just
noticed
> >> > > > them there
> >> > > > > > > the other day so passing it on to you.
> >> > > > > > >
> >> > > > > > > BrianB2.
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > > --- In amibroker@xxxxxxxxxxxxxxx, "jwmc107" wrote:
> >> > > > > > >>
> >> > > > > > >>
> >> > > > > > >>
> >> > > > > > >> Is there any AFL code to scan for Bulkowski's chart
> >> > > > patterns and
> >> > > > > > > to
> >> > > > > > >> share with?
> >> > > > > > >>
> >> > > > > > >> Thanks,
> >> > > > > > >> Jim
> >> > > > > > >>
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > > Please note that this group is for discussion between
> >> > > > users
> >> > > > only.
> >> > > > > > >
> >> > > > > > > To get support from AmiBroker please send an e-mail
> >> > > > directly
> >> > > > to
> >> > > > > > > SUPPORT {at} amibroker.com
> >> > > > > > >
> >> > > > > > > For other support material please check also:
> >> > > > > > > http://www.amibroker.com/support.html
> >> > > > > > >
> >> > > > > > >
> >> > > > > > > Yahoo! Groups Links
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > >
> >> > > > > >
> >> > > > > >
> >> > > > >
> >> > > >
> >> > > >
> >> > > >
> >> > > >
> >> > >
> >> >
> >> >
> >> >
> >> >
> >>
> >
> >
> >
> >
> >
> > Please note that this group is for discussion between users only.
> >
> > To get support from AmiBroker please send an e-mail directly to
> > SUPPORT {at} amibroker.com
> >
> > For other support material please check also:
> > http://www.amibroker.com/support.html
> >
> >
> > Yahoo! Groups Links
> >
> >
> >
> >
> >
> >
> >
> > --
> > No virus found in this incoming message.
> > Checked by AVG Free Edition.
> > Version: 7.1.408 / Virus Database: 268.13.9/490 - Release Date:
10/20/2006
> >
> >
>
Content-Description: "AVG certification"
No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.1.409 / Virus Database: 268.13.27/517 - Release Date: 11/3/2006
|