|
PureBytes Links
Trading Reference Links
|
this should be very interesting thread to the minority of the list
participants that are systematic traders and rocket scientists.
i figured that it would be the ultimate market study oriented thread:
*** let us in short go over the most common and not so common
<<market modeling techniques>> ( forecasting techniques ), their basic
math, classify them, go over their advantages and disadvantages,
give them 1-10 stars as far as the success rate in term of utilizing those
in your trading system...
sounds like fun, aint it?
for those who are lost here is a little introduction:
to be able to come up with a successful tradable system that has
high % winners coupled with high win/loss ratio and other criteria
you need to model the markets that is to come up with a *mathematical
model that describes the theory price movements thru a *math
or what we call the *process.
once a suitable mathematical model is found and the process math is worked
out, we go into *phase two which is working with the model by utilizing
the math to *forecast prices and *make decisions base on those forecasts.
as the end of the result we come up with
*buy and sell arrows on the tradestation screen that tell us when to enter
and when to exit. ( fail to go thru the above process and you will have a
clusterf...k of trading signals )
in short the PATH is: ( whether you realize it or not )
market theory ( why prices move and how they move? ).... then
math model ( process, what is the math behind price movement )....then
forecasting ( where the price will go based on the math model and
input parameter ? )... then
decision making ( do we buy or do we sell based on the forecast )
evaluate the success rate ( how good is the model and is this model
tradable? )
ok?
now everyone of us tried different approaches starting from the simple ones
like
pattern probabilistic models to more complex once like lately mentioned
state
space model or NN models, etc...
most of traders have no good models and therefore have to good trading
systems :-)
on top of that...
some of us have been looking for years for that perfect model that explains
it all
that we call the holy grail of system trading and some of us are just
starting to
get into it... so this might be of interest
SO....
the purpose of this thread is to share the experience thru discussing
different
types of market modeling techniques, the process math descriptions,
forecasting
methods based on those models, literature references, advantages and
disadvantages and the success rate based on utilizing those in trading.
( this does not have to be applied only to trading systems, could applicable
to options trading, derivatives pricing, non price time series forecasting,
etc.. )
these would be an equivalent of the encyclopedia of market modeling
techniques compiled by us... or at lleast a compilation of the existing
pointers to
market theories.
if you are a smart person, then you must realize that this just might well
be the
ultimate thread... the review, so to speak, of the search for the holy
grail...
so i encourage you to participate as much as you can.
this might save us some months of work and jump start our research since we
can filter out junk from something that works, bounce ideas, etc...
here is the framework for this thread ( PLEASE STAY ON IT ):
1. Give the name of the model or process ( ARIMA or Random Walk for
instance )
2. Classify it ( Stochastic, Non Probabilistic, Deterministic,
Chaos NLD, General Time Series Fit Model, Special, NN/AI, New etc )
be specific about classification ie.. if it's mix then describe it as
such...
3. Give basic math ( if you know it ), give the field of math that the model
utilizes or at least the general formula for the process OR
4. Book or publication reference to it. ( Box Jenkins, etc. )
5. Tell your story as far as playing with the model and testing it.
6. Describe your general view about the model, its advantages and
disadvantages
7. Since we are TS oriented tell us if you found implementation of the model
in EL, VB, C or other code.
8. Tell the list the success rate of it if you have incorporated it into
your system.
9. Give your overall score from 1 to 10 stars... 2 and 1/2 stars...
10. Tells us if you are an expert in some of the modeling techiniques or
at least tell the list your current points of interest or some new stuff
that you found...
let's see if we could pull that off on the this or omega list and bounce
some ideas.
we should at least find some individuals that have some experience to
contribute.
my personal interest in this is to bounce ideas and see where people go
and may be give and get a couple of pointers myself as far as
where i found the success and failure as far as different approaches...
***if you have no interest in this thread then don't divert this thread into
discussing
1st grade algebra or what computer you need to run your trading software,
please...
( or better yet Mark Brown, why don't you create a list called HolyGrail
and maybe we can take in
some people and off load some traffic onto that, how's that? )
also, there are hundreds of market models out there so the job is to filter
out what
works or might work from what did not work, obviously will not work or might
work
if altered...
the best way is to start with the most known one such as random walk theory
or probabilistic models ( indicator or price or combination patterns ),
how's that?
take one model apart and move on...
sounds like ultimate fun to me....
bilo.
ps
if there is enough interest*
in the next post i will give general classifications and the descriptions of
the models
but first i would like to see some hands as far as who wants to be
the millionaire system trader :-) and is interested in this thread.
|