|
PureBytes Links
Trading Reference Links
|
Having just done a 2-hour lecture on Data Mining and Market Forecasting
at the Business School last night, I just HAVE to throw in my comments:
> > For many years now Analysis techniques have been evaluated from traders
> all over the globe with the intent of creating a better system of making
> money.
>
> maybe ... or ... what you see in the public, both for sale and free ones,
> are not what real research is all about.
>
IMHO, the REAL research effort should be to discover and develop the
"physics" of the market --- those laws which are generally applicable
to trading markets in general. What most current systems have is
merely a set of "Rules" which cover specific situations, without any
deep understanding of WHY the Rules work --- and a great sense of
mystery when then fail ...
We are, IMHO, trading as if we were astronomers in the days before
glass lenses were developed. That we can now do it electronically only
means we can make our interpretations be in effect (right or wrong)
faster and at less cost.
> Real research in time series, pattern recognition, etc. require higher
> understanding of statistics, computer science, etc. but most people
> who claims to know something about TA are people who lack this type of
> knowledge and tools. In the beginning of the born of TA, yes, you can
I have 40 years in the saddle with time series etc. When I approached
TA in 1995, I did some simple testing of the various indicators using
time series principles. MANY of the TA IMPLEMENTATIONS (not the theory)
were laughably, frightenly WRONG and unstable.
Another point is that many of the TA algorithms rely purely on
"looking over your shoulder" historical data (not a professional
signal analytic term :=}}} ), when better, more stable results
should (untested yet) be obtained by going back and re-calculating
the TA indicators after what-would-have-been-future data arrives.
> have someone with average knowledge to observe some simple concepts
> and their recurring patterns, but when these basic knowledge get discounted
> by public understanding, the more complex concepts are not easily
> discovered any more.
>
> > Many traders who began trading manually later evolved into electronic
> trading because we were told that it was better.However we as traders still
> must monitor the markets,make accurate decisions to protect our investments
> and also look for formations and read live news so we as traders can
> "predict" the markets.However I have seen no software that can do this,that
> can actually take place of a trader... I
>
> depending on the time frame you are talking about, arbitration of future
> index and the
> corresponding basket of stocks can be considered as a form of auto system.
> It is a
> clear and simple example for normal souls what advancement in TA can do.
> and still
> I suspect most people still do not understand what buy sell programs means
> in
> index futures - especially we have CNBC type of publicity of it :)
>
> > also know that there are no companies claiming that their product can
> replace a trader.There software only confirms what a trader already
> knows.So what am i getting at well it's simple to me there is really no
> such thing as electronic trading however there are traders and there is
> trading and technology in regards to helping traders have not came very
> far.
>
> Those who is making money keep the trading secrets to themselves.
> Those who cannot make money in trading sell what they have develop.
> Those who cannot pull the trigger keep b.s. in the public fo how great they
> are.
> plus, a new bleed - those who do not trade nor develop new TA stuff sell
> tools
> for constructing systems.
"Those who CAN, DO."
"Those who can't, teach." (or preach, in the current situation)
>
> -Lawrence Chan
> p.s. so all great trading techniques are rediscovered again and again
> by chance, by hardwork, or by luck, thats all it is :)
>
Having proven (to my own satisfaction) that there IS a "physics" to
the market --- and having developed a modest set of Rules and Indicators
that predict short-term market movements (http://apk.net/~lake), I'm
about to "properly" implement some TA indicators and see if they work
as well, better, or worse than my Rules. This will be part of
re-engineering my system to run in real-time , hopefully
on every tick.
No bragging (can't make a living off the system yet) and no promises.
Cheers,
Rob Lake
rbl@xxxxxxxxxxx
>
|