|
PureBytes Links
Trading Reference Links
|
Chuck,
Here is European site, which used to be free at one time
and I used to be surf there so often in the past.
http://www.market-topology.com
Click on the demo (green)
U will get correlations and anti-correlations. They
also provide a topology tree map of densily related clusters but
I think this is only available upon subscription. The tree map
is more revealing than the correlations. Their correlations
are more reliable though cos they use long term data.
My interest is in developing tree maps on my own using SSA which is
an analysis of smallest "distance" between pairs, where distance
could be defined in any manner. U will find this under "multi-
dimensional scaling" in statistics books.
At the website u will see that currently, CSCO is very strongly
related to CSN (Amex). What is CSN, I don't know, but plot these
together in Yahoo Finance site and u will see CSN almost always
pulls up CSCO. Do u know what is CSN and why it is so correlated
with CSCO?
Anyway, I'm interested in my own tree maps and I think I know
better ways to do it for shorter time frames - atleast I will
try.
nand
--- In amibroker@xxxxxxxxxxxxxxx, "Chuck Rademacher"
<chuck_rademacher@x> wrote:
> Where can I go to learn more about SSA?
>
> By the way... I did say 8,000+ stocks, but reality there would be
far fewer.
> I do pass all 8,000+ through my filters, but I'm only interested in
stocks
> above a certain price and with a minimum turnover, etc.
> -----Original Message-----
> From: nkis22 [mailto:nkishor@x...]
> Sent: Tuesday, April 29, 2003 5:48 PM
> To: amibroker@xxxxxxxxxxxxxxx
> Subject: [amibroker] Re: (More) Real World Trading
>
>
> Hi Chuck,
>
> I have often successfully traded one sided DIA-options using the
> same pairs idea. The pair I use comes from The Dow Theory
> that specifies the relation between TANSPORTATION and DJiA
>
> The critical thing is that the correlation between any
> pair must be there atleast in theory or some solid reasoning
> as to why the pair ought to be correlated. With 8000 stocks
> there would be about 64 million correlated pairs, and some
> very stong correlations will appear just by chance.
>
> With smaller sets of stocks, I have looked at relatedness
> via SSA ( a type of Euclidean distance) analysis that
> will show different clusters. I am going in this direction
> but have not developed a system or even done a single trade as
> yet. I think SSA is better than looking at correlations which
> can be "dishonest" or "spurious" and quite there by chnace. Give
> us an example, may age and the price of gasoline is extremely
> highly correlated but meaningless!
>
> Interested in your work very much.
>
> tia
> nand
>
>
>
> --- In amibroker@xxxxxxxxxxxxxxx, "Chuck Rademacher"
> <chuck_rademacher@x> wrote:
> > You may recall that I have been making a concerted effort to
come
> up with
> > some system ideas using "pairs trading". I don't know how
> successful any
> > of you have been at coming up with systems for shorting stocks,
but
> I am
> > extremely frustrated by the need to maintain a dollar and/or
beta
> neutral
> > portfolio in keeping with the rules of the funds that I
manage. As
> a rule,
> > my short trades are lucky to break even. Fortunately, my long
> trades are
> > quite profitable.
> >
> > I am going to put forth some ideas that should have been more
> obvious to me
> > some time ago and may already be obvious to you. Like a lot of
> you, I have
> > what I think are excellent systems for trading the long side. I
> also have
> > excellent systems for trading "pairs". My pairs trading has
been
> 100%
> > isolated from the other styles of trading that I do.
> >
> > My daily search for pairs to trade never really had a starting
point
> > (watchlist). The system takes each of the currently active
stocks
> (8,000+)
> > and tries to find another stock to trade against it. There
was no
> starting
> > bias. The system looks for the most highly correlated pairs of
> stocks to
> > trade; one long and one short.
> >
> > Late yesterday, I had a thought come to me to take the just the
buy
> signals
> > generated by my best "long" system and only look for a pair to
> trade against
> > each of those. It's early days in the research, but it looks
like
> an
> > exceptionally good idea. Instead of making a mediocre 1.5 to
2%
> per month,
> > the research suggests almost doubling that return while still
> maintaining a
> > dollar, beta, sector and volatility neutral stance.
> >
> > Any comments would be appreciated and I'd be happy to work
closely
> with
> > anyone interested in pursuing this endeavour. AmiBroker is
> definitely the
> > tool of choice for this job, hence my posting to this group.
> >
> > Cheers
>
>
> Yahoo! Groups Sponsor
>
>
>
> Send BUG REPORTS to bugs@xxxx
> Send SUGGESTIONS to suggest@xxxx
> -----------------------------------------
> Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
> (Web page: http://groups.yahoo.com/group/amiquote/messages/)
> --------------------------------------------
> Check group FAQ at:
> http://groups.yahoo.com/group/amibroker/files/groupfaq.html
>
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of
Service.
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get A Free Psychic Reading! Your Online Answer To Life's Important Questions.
http://us.click.yahoo.com/O10svD/Me7FAA/AG3JAA/GHeqlB/TM
---------------------------------------------------------------------~->
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
|