|
PureBytes Links
Trading Reference Links
|
I have been using a powerful emerging product, called PowerST,
http://www.powertesting.com, for almost a year now. I sent an email to this
list on February 28, describing some of its capabilities. I will repost
that or forward it to individuals who would like to see it.
With respect to the issues in this email thread, PowerST implements a
trading platform that gives the user control over whether or not historical
testing should take signals on locked limit days. In addition, it is
designed to allow users to customize their trading systems, portfolio and
money management analyses, using C++ (or C). In particular, there is
readily available free or inexpensive C and C++ source code that can be
integrated into PowerST trading systems. I'm not suggesting that the kind
of power that packages like Mathematica provides is available cheap.
However, there is a lot of useful statistical software available free or
inexpensively, and it is easy for a programmer to integrate it into PowerST
trading systems.
Paul Zislis
----- Original Message -----
From: "brian" <polar@xxxxxxxxxx>
To: <omega-list@xxxxxxxxxx>
Sent: Tuesday, April 03, 2001 2:24 AM
Subject: Re: Why futures truth doesn't care about stealing your system
> I have asked in the past if anyone has ever tried mathematica's technical
trader package. I used the basic mathematica package years ago in school
and it's very similar to matlab in its math and graphics capabilities. The
problem with
> moving to a completely "open" platform is that you have to build
everything from scratch with basically no technical support. I don't mean
just buy/sell coding but more difficult but also important aspects like not
allowing buy/sell entry on
> limit up/down days, etc. Someone who trades with their software will
usually be responsive to making these types of changes. However, I doubt
the matlab/mathematica people will have a clue if you call for support
asking why you're getting
> buy signals in limit lumber when it couldn't have occurred. As an aside,
I completely agree that those packages do provide better all around number
crunching and graphics - and they are portable across platforms. But the
problem as I see it
> is that there is no main framework established to work with. While I can
programming is fair I prefer to use the Japanese approach and customize and
upgrade an existing product rather than reinvent the wheel myself. If at
some point these
> companies begin to develop software for trading system design I would
happily pay for their product as I am tired of importing/exporting data
across multiple applications but I don't see it occurring anytime soon. I
spoke with the sales rep
> at mathematica a while back and when I asked simple questions as to what
the technical trader package could do he seemed baffled. Many places do use
these packages in their risk management side as they can quickly calculate
var/bp values but
> I don't know anyone who uses these packages exclusively for system design.
I do know one person who developed his testing platform with APL and Gauss
but he was somthing of a programming prodigy and it still took him close to
two years to
> work the bugs out. Maybe if enough people call one of the major math
computing companies will decide to offer something for traders but since the
universe of technically proficient ssytem developers is relatively small
compared to the
> number of university research research types ) I still have my bets on an
outside trading company/programmer developing a package for these products
first, if at ever.
>
> > Subject: Re: Why futures truth doesn't care about stealing your system
> > Date: Mon, 2 Apr 2001 00:05:12 -0400 (EDT)
> > From: Lawrence Chan <stnahc@xxxxxxxx>
> > To: omega-list@xxxxxxxxxx
> >
> > well, OR looks like failed in the software
> > business and have to remake itself into a brokerage.
> >
> > maybe they priced their stuff too cheap :)
> > the prices you see for matlab is institutional
> > pricing - that's why they look expensive.
> >
> > high power tools are never available to
> > the public anyway ...
> >
> > -Lawrence
> >
> > --- Robert <robert_h@xxxxxxx> wrote:
> > > Every now and then someone says MatLab is good for
> > > financial programmers.
> > > Look at the prices below from the MathWorks web
> > > site. They make
> > > TradeStation at $2500 look cheep. Just the C++ Math
> > > Library is $900! Their
> > > "Real-Time Workshop" is $21,870!! The MatLab stuff
> > > is for people with more
> > > money than brains.
> > >
> >
> > =====
> > Lawrence Chan http://www.tickquest.com
> > Home of trading tools NeoBreadth and NeoTicker series
> >
>
------------------------------------------------------------------------
>
>
|