PureBytes Links
Trading Reference Links
|
I'd add just a few caveats to the essential point that Trader J and Bob
Fulks made. The caveat speaks to the objection that Tim Morge made in a
separate reply.
The caveat is simply this: I'm not sure that "programming skills" per se
provide a fundamental advantage over time. Equally, if not more, important
is one's ability to observe market behavior and to develop (or borrow from
reading) a theory of that behavior that can be designed into a program
concerned with trading. There are alot of "programmers" out there who won't
ever be successful because they don't take the time to observe market
behavior and/or develop their own coherent set of theories about the market.
I think this is close to the point Tim Morge is making. He's saying that
there is knowledge about market behavior that a "mere programmer" can't
compete against. I agree with him on this. I believe though that Tim
overstates the case by underestimating the extent to which "knowledge"
(including his knowledge about market behavior) can be captured through
robust software analysis and design techniques and re-expressed in the
language of a computer program. This is the point that Bob Fulks makes re:
the knowledge required for chip design.
I would take Bob's point a step further. In institutional settings, over
time, even mediocre "mere programmers" as a group may be very competitive
given the longevity of their trading program work product.
For individuals rolling their own, I think the situation is different. To
me, the knowledge that one must have to build a trading system that has odds
that I feel comfortable with is intimidating. But there is an enormous
amount of market knowledge in the public domain that one can incorporate
into one's consciousness through mere reading. So, my proposition (hope?)
is this: that the individual programmer/traders who become successful most
rapidly will be readers....
Steven Buss
Walnut Creek, CA
sbuss@xxxxxxxxxxx
"There's nothing more practical than good theory."
-----Original Message-----
From: Bob Fulks <bfulks@xxxxxxxxxxxx>
To: Trader J <editorial@xxxxxxxxxxxxx>
Cc: Neal T. Weintraub <thevindicator@xxxxxxxxxxx>; Steven Buss
<sbuss@xxxxxxxxxxx>; omega-list@xxxxxxxxxx <omega-list@xxxxxxxxxx>
Date: Monday, August 03, 1998 6:23 PM
Subject: Re: Trading as a way to financial success (a reply)
>At 7:28 PM -0400 8/3/98, Trader J, or Omega Man, (or whatever) wrote:
>
>>I tell young traders: Work on your automation/systems. Don't buy mine.
>>Build your own. Know which way you want to trade - then automate
>>everything. Decide what you want the machine to do, and make it work for
>>you. When it becomes entirely machine against machine, the best
programmer
>>wins.
>
>You got it man!
>
>I worked in the field of designing software tools used by designers of
>microchips. For many years, a really good chip designer could run rings
>around the best software systems. They were "creative", had "intuition",
>and could see things that the machine couldn't see. Then it was
>neck-and-neck for a while as the humans fought to stay ahead.
>
>Then, the software moved ahead and now there is no looking back. No company
>would ever consider designing a complex chip by hand anymore. Now, software
>will win by at least an order of magnitude.
>
>The same happened with mechanical design.
>
>It is only a matter of time with trading. Some of the brightest software
>guys I had quit and went to work on Wall Street where they were paid
>ridiculous sums. They are given supercomputers to play with and treated
>like gurus.
>
>The key is that software can continue to accumulate knowledge over time.
>Each new person who works on a program adds new features to what already
>works. It only gets better over time. Whereas people retire and die and
>take their secrets to the grave with them, and each new generation largely
>has to start over.
>
>Why do you think there is always such a shortage of computer programmers?
>
>"When it becomes entirely machine against machine, <and it will soon>, the
>best programmer wins."
>
>Bob Fulks
>
>
>
|