Advanced Automatic Investor

[aptus]

Hi Conrad,

Very interesting suggestion. I had played around with self-adjusting SAFEs over a year ago, but without much success -- although I did have high hopes for it and even coined a name. I called it, "IntelliTrend."

However the results were consistently dismal and I chalked it up to trying to adjust a long-term parameter (i.e. SAFE) using short-term variables (i.e. daily, weekly or monthly price updates).

After that, I theorized about an algorithm to dynamically adjust parameters based on a longer history of price data using some Artificial Intelligence techniques. However I never did implement anything in order to test it.

Your suggestion of updating the portfolio's start state after each price update is interesting in that it has the potential to work. However I think there would be quite a few issues you'd have to resolve before you'd be able to implement it.

Here are some off the top of my head:

1. I don't think you can use 30 days to build the Model. Remember that AI is a long-term strategy and using only 30 days worth of data would not make a good Model given AI's algorithm.

2. I think rebalancing gives its greatest advantage when done on a 2 or 3 year basis (perhaps even annually, but certainly not on a daily, weekly or monthly basis).

3. You might have to come up with additional cash to deposit into the Portfolio at any given time. How would you handle it? Would it be initially part of the portfolio (but kept in a separate "partition") or would you expect to have it available from an external source when necessary?

If you do experiment with this, please let me know your results.

I think a good first step is to test any modification over the Nasdaq 100 for 3, 5 and 7 years (ending with the current date) and compare it to a standard (such as AIM or Buy and Hold).

Then try the Dow 30 over the same periods. Then try both of these over various time-periods (such as 1969 to 1979 etc.)

If the results are good, then it's time to really start testing. If the results are not good, it's time to look for a new strategy.

Regards,
Mark.

http://www.automaticinvestor.com