InvestorsHub Logo

aptus

09/16/02 1:50 AM

#218 RE: jackpark #204

Hello Jack,

Very nice to hear from you. When I started writing Automatic Investor, a few years ago, I spent a great deal of time at your thinkalong site.

I was especially interested in your learning AIM.

This board is meant as a forum to discuss general AIM improvements (some improvements might end up in Automatic Investor) rather than a forum for a particular implementation of AIM. However if you have ideas from your open source Java AIM engine, please feel free to bounce them around here.

I think there's an abundance of very competent AIM tweakers reading this board.

I'm looking forward to reading about your ideas.


Regards,
Mark

http://www.automaticinvestor.com

Conrad

09/18/02 5:36 AM

#244 RE: jackpark #204

Hi Jack...No, I am not taking you to Cuba!

I did a lot of reading on you Think Along Site that Tom Veale linked me too earlier this year. I was impressed with your ideas as it reminded me of self-adjusting algorithms that are used in industrial process control computers. I did react to you site but I believe it was dormant at that time.

Your message interests me very much. I have designed an AIM Variant(VORTEX) and it was programmed by a programmer that uses C++(I think). The point of my contacting you is that I am looking for an optimisation package or design process to optimise VORTEX. Currently Vortex has 5 active(programmed) parameters while two more are being planned.

My optimisation objectives have been discussed on this Board, and I have had discussions, among others, on this with Mark Hing from AI. My objective is to define a variable share price history(I could for example choose 1 month or 1 year, etc) and run a optimisation with that data to set the parameters for the stock. Then I would periodically re-optimise as I thought it necessary, based on market behaviour. An option would be, of course, to link that to VORTEX itself and then the common DataBase would provide the stock prices to do the optimisation on the previous price history. One can even make the optimisation an automatic feature based of predetermined price behaviour(breakouts). These latter points are less important right now. A point of interest of mine is the issue of modifying the share price history prior to feeding it to the optimisation module. I would do that but I am not quite satisfied if it is worth the trouble from a practical point of view. Things like that I need to work out yet. The option would be to smooth the price curve and transform it into an analytic function. I see pro's and con's on this. As you see, there is plenty to Think Along on this Vortex objective!

My main barrier for progress is that the actual know-how for the writing an optimisation program is not (yet) our field of experience. Your experience or your system may be something I am looking for. Mark Hing has advised me towards some ideas as well. Your system interests me. Is it for sale or are you offering your services in this direction?

For your information, I have just recently translated VORTEX to English and you can download a free demo(full version) to see the program interface and functioning:

http://www.vortexcw.nl ------ > Link: Vortex Method------ > Contents: Stock Market------ > Download.

Here and there some Dutch text may still appear or the translation may not yet be in line with US investment jargon. This requires a little maturing!

I like to hear from you if your know-how can be applied to Vortex optimisation, and if so, at what terms and conditions?

Regards,



Conrad