AW: testing software- build or buy??

["Volker Knapp" <vk@xxxxxxxxxxxx>]

Hi,

why not write the logic and test it on a portfolio of stocks/futures/mutual
funs or what so ever inclusive

bet-sizing,
risk-control stops,
bad-tick filters,
adaptive parameter settings, etc.,

required to convert a good system idea into a robust "tradeable" trading
system.

I don't really like it if people just test a system on one market. But that
is my personal preference. But then of course you have to decide what
software to use for trading it - or just use the same software.

I just feel that TS user are so limited in there possibilities of testing a
system, because they always have to test one future after the other, and
except it as if that is it....

But then again I am biased, and will stay biased until TS has improved its
product.

Volker Knapp


-----Ursprungliche Nachricht-----
Von: Bob Fulks [mailto:bfulks@xxxxxxxxxxxx]
Gesendet: Mittwoch, 6. Februar 2002 17:08
An: Marc Lawrence
Cc: omega-list@xxxxxxxxxx
Betreff: Re: testing software- build or buy??

I would suggest getting some EasyLanguage programmer to test the
system idea in TS 4.0 or TS 2000i to see if it really works. This is
usually a pretty simple task for most system ideas (unless you are
using complex mathematics).

Then, if it really works as you expect, figure out what software
platform you want to use for actually trading it.

You can then get someone to program the system in the new platform,
add the bet-sizing, risk-control stops, bad-tick filters, adaptive
parameter settings, etc., required to convert a good system idea into
a robust "tradeable" trading system.

Bob Fulks


At 4:52 PM -0800 2/5/02, Marc Lawrence wrote:

>Hello all,
>
>I would like to test a couple of systems and have been
>concerned that:
>
>1. Tradestation is not supporting the older versions
>and is going to a broker model-  their focus is not on
>building testing software anymore.
>
>2.  Is it worth getting an older version TS4 or 2000i
>and learning "easy Language" only to find it can't do
>what you had in mind.
>
>or
>
>Hire a programer to build a database to do this type
>of analysis.  Does anyone have experience with this?
>The one thing I like about this approach is you are
>not at the mercy of Tradestation and can trust the
>results-  although it is probably going to be a
>sizeable undertaking.
>
>Any comments would be appreciated -  thanks-  Marc