Replies to post #3688 on A.I.M. Users Bulletin Board (AIMUSERS)

[whitelake]

Hey Tim, It's probably going to be kind a dead around here today but you can check out Tom's great page http://www.aim-users.com/ if you haven't already. Tom is better than a back test. He is a living forward test as you will see.



whitelake

[aptus]

Hello Tim,

I am almost certain that nobody has done what you suggest, so the door is wide open for you.

My preferred method of testing AIM is to, as you say, turn the AIM knobs. However I do it in a very methodical, rather than in an ad hoc, way.

I've found this to be the best way of determining what works and what doesn't.

Theoretically, your method of a "mathematical formalism that relates the inputs, engine and goals" would be better -- if you can do it.

However you might find that there are too many variables involved to come up with such a formalism for more than just specific cases. If you do it, however, I'd be most interested in seeing it.

I'll look for your future results.

Regards,
Mark.

http://www.automaticinvestor.com

[OldAIMGuy]

Hi Tim and Welcome,

For me it was a philosophical decision to use AIM. I'd tried other things with some levels of success, but didn't like the extra risk exposure and time consumption needed just to do a bit better.

AIM also hoards cash as I had been doing on my own before I found AIM. This gave me a reason to like it right away. <g>

I've used both methods of working with AIM. Selecting AIM characterized stocks for use and also adjusting AIM to work with stocks that don't necessarily fit perfectly into the AIM method.

Since long term total return can come from three things as an AIM investor, we can pick and choose according to what satisfies our personal goals. Price fluctuation drives AIM. Long term upward price bias is also good for total return. Finally accumulated dividends add their share. So, part of the investors selection process has to do with their own stage in life and what's most important.

My personal goals include capital preservation, income and growth. I want to meet or beat inflation, I need income for living expenses and I don't want to indulge in strategies that might have severe capital bashings if I'm not doing things right.

I think Mark Hing may have done some of the type of analysis you suggest. Yes, I feel that my long term use of AIM isn't to "get rich quicker" but more to help assure that I can stay in the game longer.

Best regards, Tom

[Rien]

Hi Tim,

I am rather busy, and have thus missed the discussion so far on the testing issue.

Seems to me you are proposing to create "artificial" stocks and then try different algorithms (AIM, XDEV) on them. If so, then we might be interested in the same thing.

I am currently writing a java application to do this kind of testing. The idea is to have a test-bed in place that can easily be modified (via programming-compiling) for new algorithms or enhancements on current algorithms. I am using this application as a learning example. I.e. I wanted to learn java/swing and use this example to have some hands-on experience. So far I am progressing faster than anticipated, but lots of stuff still needs to be done. I only started recently (about 2 weeks ago), and have to do it in time-sharing fashion as I also have some other projects running. If everything goes as planned, I should have something ready by the end of summer.

As to the artificial stocks, I wanted to include an option to let the user define stockprices by setting:
1. A start offset
2. A slope (i.e. a second increasing or decreasing offset)
3. A couple of sine-waves (5) of which the user can define amplitude, period and start-phase.

I know that 2 is in fact a kind of sine wave, but I rather fancy the idea :-)

BTW: the application also allows to use real stock data.

Maybe we can come back to this testing in a few months.


Best,
Rien.