|
PureBytes Links
Trading Reference Links
|
Am I glad to hear I am not the only one who feels this way! Not only
is TA no Holy Grail I do not think of them as much better than
broken clocks that chime twice a day. I decided to do things my way -
teach the robot to trade like a discretionary trader. In a very
crude manner, for sure. But despite the crudeness the robot trader
(mechanical system) has an edge because of it's consistency. We
cannot measure the odds of our decision making unless we are
consistent, at the same time we do not learn efficiently without
reliable feedbacks, this I feel is the problem with discretionary
trading.
Using a combination of indicators does not necessarily improve the
resulting accuracy - a set of 5, 10 or 20 that agree, among hundreds
and thousands of broken clocks, is no more guarantee of accuracy
than one indicator alone. However "decision making" based on
multiple TA does "smooth the accuracy curve" which is important for
me. I would rather my TA accuracy swing between 45% - 55% and not
between 30% - 70%, if I am confident of that the Average Win:
Average Loss ratio will take care of the rest. Assuming 40% win the
AW:AL ratio needs to be 1.5 to break even and if through exit design
I can achieve an AW:AL ratio of 2.5 I will have some cushion for
Walk Forward degradation. The AW:AL ratio is easier to control and
degrades less than the %win ratio because it is more money
management than TA dependent.
I stick with a few simple indicators and won't use anything I do not
understand. My favorites are RSI, SRSI, Bollinger Band and
candlesticks, not because I think they are better but because I know
them better.
Sursod
--- In amibroker@xxxxxxxxxxxxxxx, "brpnw1" <tradermail@xxx> wrote:
>
> I feel I led some people here astray. I want to say that no single
> indicator may be used to predict the market 100% of the time. If
it
> existed, we would know about it, because it would be our holy
grail
> and it's all we would talk about here.
>
> What I have been trying to say is that a number of different TA
> systems are required to make an "uber" system that contains them
> all. The "uber" system contains decision-making code that taps
into
> such things as market breadth, open interest, and trend detection -
-
> all for the purpose of emulating the human decision-making process
> regarding which technical indicators to use.
>
> The key is not in the depth of the indicator code complexity. It's
> in emulating the human decision-making process regarding which
> indicators to use at any given time.
>
> There should even, perhaps, be a separate AFL library or library
> section that focuses strictly on code that emulates the decision-
> making process. It would do everyone well to stop pursuing the
math
> and engineering skills, and begin pursuing the ability to break
down
> the decision-making process regarding which TA indicator
> combinations to use for different situations.
>
> After testing thousands of indicators and combinations of
indicators
> over the last 6 years, I can now say my best indicators are the
> simplest ones, some of which are ones that were handed to me in my
> first month or two of playing with Amibroker. The hard part has
been
> in getting the trading experience under my belt so that I can code
> the decision-making process mentioned above.
>
> Thanks for reading,
>
> Namaste,
>
> ~Bman
>
> --- In amibroker@xxxxxxxxxxxxxxx, "cstrader" <cstrader232@> wrote:
> >
> > Although I very much hate to say it, I am indeed skeptical that
> there are
> > any technical systems that work consistently. I am also
convinced
> (see Ly's
> > argument) that if technical analysis ever did work, it works
less
> well now
> > than it did in the past.
> >
> > I would like to hear from any technical trader who can provide a
> complete
> > and independently verified list (for instance on
> www.timertrac.com) of his
> > or her trades that show a profit that beats some benchmark (say
> sp500) over
> > a consecutive period of 3 recent years.
> >
> > Ly, can you explain why volatility analysis is different than
> technical
> > analysis? Can you give examples of volatility systems that
might
> be
> > useful? Also, can you explain your statement regarding
> the "correlation
> > between volumes and prices?"
> >
> > One example of a volatility-based system that may be successful
is
> the
> > "fasttrack" approach (see for instance
> > http://www.greenmountainaccess.net/~wwgansz/.
> >
> > really enjoying the thread!
> >
> > chuck
> >
> >
> > ----- Original Message -----
> > From: "Tom Tom" <michel_b_g@>
> > To: <amibroker@xxxxxxxxxxxxxxx>
> > Sent: Saturday, December 02, 2006 12:50 PM
> > Subject: Re: [amibroker] Re: Random Walk - step 2 - :
> Predicitable ?
> >
> >
> > > Hi Bman,
> > >
> > > Sure there is psychological and human behaviour in this game,
> and it has
> > > to
> > > be considered.
> > >
> > > But the financial instition should say us "yes it is
> predictable"... so we
> > > put all our money on the market for them. If they say, "it is
> random walk"
> > > people will leave the market and give less money to it.
> > > It need to be balanced i think...maybe yes maybe not, so
mystery
> is keep
> > > and
> > > financial institution have maximum cards to play in their hand.
> > >
> > > I aggree 100% with Chuck about this line "technical analysis
has
> not been
> > > validated in controlled studies"...
> > > It is true, i have never read (if someone know where to find,
i
> am very
> > > interrested, thx) a clean scientific demonstration about
winning
> trading
> > > system... nor an old mechanical trader publish any trading
> reconstruction
> > > based on real trade winned by his trading system and showing
> precise
> > > reports
> > > and indicator used... and it frighten me sometimes, because
> maybe after
> > > all
> > > the winner we show us are only a small part of the people
which
> take a big
> > > risk and win (big risk = big return if lucky = good trader ?).
> Those who
> > > take a big risk and did'nt win are no more here.
> > > Statically, on all the trader over the world, their is some
who
> can be
> > > lucky
> > > and win 10 years , 20 or more consecutive years... few
people...
> but
> > > possible. Are their technics consistent ? Do they adapt their
> technics
> > > over
> > > the time ? (so profit cannot be consistent because we cannot
for
> sure have
> > > a
> > > good trading system everytime).
> > > Why not a book on a big trading looser ? : )) so trader (bad
or
> good)
> > > would
> > > make money not by trading but by publishing book héhé.
> > > YES we can make money on the market it is a fact, but we have
to
> be
> > > very...
> > > very... very carrefull i think if we want it to be consitent
> over the
> > > time.
> > > The hard compromises we face is : Commission / Returns and
> Risk / Profit
> > > expected.
> > >
> > >
> > > Their is are two book on the subject, i find the title funny :
> > >
> > > 1- A random walk down wall street, by Burton G. Malkiel
> > > http://people.brandeis.edu/~yanzp/Study%20Notes/A%20Random%
> 20Walk%20down%20Wall%20Street.pdf
> > >
> > > 2- A non-random walk down wall street, bu Andrew W.Lo and A.
> Craig
> > > MacKinlay
> > > http://www.amazon.com/Non-Random-Walk-Down-Wall-
> Street/dp/0691092567
> > >
> > > It show well the problem.
> > > I did'nt read the first one (just the abstract)
> > > I just read fastly thez second one. Very good, go deep in the
> problem with
> > > mathematic backgound to show assumption which are made inside.
> > >
> > > First one say from its abstarct : "this is random walk, and
all
> that we
> > > can
> > > do is good managing of risk"
> > > Second one say : "this is not random walk because volatility
> don't follow
> > > random walk model"
> > > All seems about volatility :
> > > First one : risk managment = manage portfolio gievn the
> volatility
> > > (=risk).
> > > Second one : volatility is not random
> > >
> > > So to go deep on the subject :
> > > Does someones here make pure volatility based trading system
on
> Amibroker
> > > ?
> > > Can we have his feeling about that ?
> > >
> > > Cheers,
> > > Mich.
> > >
> > >
> > > ----- Original Message -----
> > > From: brpnw1
> > > To: amibroker@xxxxxxxxxxxxxxx
> > > Sent: Saturday, December 02, 2006 5:36 PM
> > > Subject: [amibroker] Re: Random Walk - step 2 - :
Predicitable ?
> > >
> > >
> > > The fact that people make consistent money off the stock
market
> is
> > > evidence that the markets are not random. It appears that self-
> > > purported "experts" who likely work for large financial firms
> will
> > > go to great lengths to use data to help people forget that the
> > > markets are not random -- of course it's not random, because
> people
> > > are making consistent wins off the market, using technical
> analysis.
> > > People such as John Ehlers, for example, who have created
black
> box
> > > mehods that will always profit from the market, without any
human
> > > intervention.
> > >
> > > These financial firms have everything to gain by demonstrating
> that
> > > technical analysis is an illusion. They want to handle your
> money so
> > > they can make their profits. Don't ever believe them. They
want
> you
> > > to ride out the long-term dips in the market without ever
moving
> > > your money. They make more money if you don't move your money.
> The
> > > compliance portion of the financial industry goes to greath
> lengths
> > > to make sure that once they have your money, very few people
in
> the
> > > financial world can actually use technical analysis to make you
> > > regular profits. Try getting a job as a financial planner,
based
> on
> > > your ability to make people money using technical analysis --
> you'll
> > > never get near a desk at any firm. They don't want you to
> contradict
> > > the BS that they feed the masses.
> > >
> > > In order for financial firms to make money off you, they have
to
> > > make you lose money. Somebody always loses in the stock
market.
> They
> > > just want to make sure it's you.
> > >
> > > So continue to seek out technical analysis to make consistent
> gains
> > > in the market. Regularly read articles written by people who
are
> > > already doing this successfully, so you don't lose track of
> reality,
> > > since the financial firms are rich enough to produce a very
> > > convincing BS argument.
> > >
> > > ~Bman
> > >
> > > --- In amibroker@xxxxxxxxxxxxxxx, "cstrader" ...> wrote:
> > >>
> > >>Hi Tom Tom:
> > >>
> > >>Yes, an interesting article. I was particularly intrigued by
this
> > > line:
> > >>
> > >>"technical analysis has not been validated in controlled
> studies "
> > >>
> > >>Is there any evidence that what we are trying to do might ever
> > > work? How
> > >>could we prove that it does?
> > >>
> > >>chuck
> > >>
> > >>----- Original Message ----- From: "Tom Tom" ...>
> > >>To: <amibroker@xxxxxxxxxxxxxxx>
> > >>Sent: Friday, December 01, 2006 5:12 PM
> > >>Subject: [amibroker] Random Walk - step 2 - : Predicitable ?
> > >>
> > >>
> > >> > To go on dicussion about random walk, nice article at the
> middle
> > > of this
> > >> > page :
> > >> >
> > >> > http://www.duke.edu/~rnau/411georw.htm
> > >> >
> > >> > Combine: Random Walk and Prediction.
> > >> > Technical analysis... usefull ? Financial information ...
> > > usefull ? Even
> > >> > illegal information (hidden to public) .. usefull ? Last one
> > > maybe.
> > >> > Others,
> > >> > humm....
> > >> > This is what about deals this article.
> > >> >
> > >> > For me, next theory could be a Chaotic Fractal Near-Random
> > > Walk... : ))
> > >> > Chaotic : because spurious peak in the data wich can
initiate
> > > further
> > >> > mouvment
> > >> > Fractal : year, month, day, hour, minute, sec... same
patterns
> > >> > Near-Random Walk : Random Walk but predictable, because i
> don't
> > > think
> > >> > price
> > >> > move randomly...
> > >> > If they move randomly... tehnical or fundamental analysis
are
> > > useless, so
> > >> > there is no mean to try to trade at all, (only to give
> > > commission to the
> > >> > broker héhé).
> > >> >
> > >> > Seriously, from this article, what seems emerging from last
> > > years, is that
> > >> > price is random walk, but volatility maybe not... It is well
> > > explained in
> > >> > the article. Arch and Garch model are mentionned.
> > >> > Someone try this on AB ? Trade based only about volatility
> > > prediction (so
> > >> > predict risk, and manage portfolio depending those
prediction
> > > about
> > >> > volatility)... and so don't bother with the price random-
> walk ?
> > >> >
> > >> >
> > >> > Cheers,
> > >> > Mich
> > >> >
> > >> > __________________________________________________________
> > >> > Les révélations de la starac 6 commentées par Jérémy!
> > >> > http://starac2006.spaces.live.com/
> > >> >
> > >> >
> > >> >
> > >> > 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 NEW RELEASE ANNOUNCEMENTS and other news always check
> DEVLOG:
> > >> > http://www.amibroker.com/devlog/
> > >> >
> > >> > For other support material please check also:
> > >> > http://www.amibroker.com/support.html
> > >> >
> > >> > Yahoo! Groups Links
> > >> >
> > >> >
> > >> >
> > >>
> > >
> > >
> > >
> > >
> > >
> > >
_________________________________________________________________
> > > Windows Live Messenger sur i-modeT : dialoguez avec vos amis
> depuis votre
> > > mobile comme sur PC ! http://mobile.live.fr/messenger/bouygues/
> > >
> > >
> > >
> > > 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 NEW RELEASE ANNOUNCEMENTS and other news always check
DEVLOG:
> > > http://www.amibroker.com/devlog/
> > >
> > > For other support material please check also:
> > > http://www.amibroker.com/support.html
> > >
> > > Yahoo! Groups Links
> > >
> > >
> > >
> >
>
Content-Description: "AVG certification"
No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.5.432 / Virus Database: 268.15.9/573 - Release Date: 12/5/2006 4:07 PM
|