Advanced Automatic Investor

[DonCarlson]

Five Year Annualized Return of 32.8% with AIM variation.

I've been doing a study of the NDX-100 and SP-500 using three AIM methods:
(Std AIM, MattMod, and MACRO AIM)
The NDX-100 study is complete and the summary tables are attached. I analyzed MSN monthly data over the past 5 years for each of the stocks, each stock started with $10,000 (80% initial investment in stock and 20% cash in reserve), and each stock was managed with the same variables. The ending average Portfolio Value (PV) for each method was then placed in the summary table.

While many variations are evaluated, the average ending PV for historically typical parameters are:

$22.3 Buy and Hold -- more than doubled the investment at the end of 5 years.
.....69% of Stocks have ending PV greater than or equal $10k

$20.0 Std. AIM -- poorer results than B/H, but still doubled the investment at the end of 5 years.
.....75% of Stocks have ending PV greater than or equal $10k
.....64% of Stocks have ending PV greater than or equal B/H

$21.3 MattMod -- poorer results than B/H, but better than Std AIM, more than doubled the investment over 5 years.
.....76% of Stocks have ending PV greater than or equal $10k
.....69% of Stocks have ending PV greater than or equal B/H

$41.3 MACRO AIM -- Twice as good as the other methods,
32.8% annualized return. The only method to consistently beat B/H.
.....88% of Stocks have ending PV greater than or equal $10k
.....74% of Stocks have ending PV greater than or equal B/H


An annualized return of 32.8% with only monthly updating.....is not bad, considering the tech heavy NDX went from low values to spectacular highs during the Y2K hay-days and then crashed back to their lows and/or went completely out of business.


STD AIM Description:

A brief description of AIM can be found at http://www.aim-users.com/aimbrief.htm
The biggest changes in each book edition of Lichello's "How to Make a Million Dollars Automatically", has to be the change in emphasis between the Equity and Cash sides of the equation. First 50/50, next 67/33 and finally 80/20. In the 4th edition Lichello re-emphasied that he preferred Monthly updating. "I did not realize at the time, however, that I had fine-tuned my program to the exact point I had always sought: an optimism balance between risk and reward. I understood it only when I saw that using a twice a month checkup could become a "bear trap." Thereafter I returned to my old tried-and-true monthly protocol --- and AIM-HI. Aim High Intensity, was born."


MattMod Description by Matt Crider:

MattMod is only a slight variation of AIM. It was derived logically and not by optimization. Some consider that MattMod corrects a small deficiency in StdAim which is that StdAim does not take Cash into account in any way - it is simply a passive resevoir of the AIM action on the invested side. However, in reality Cash is the most important aspect of investment. Consider that no percentage of stock gain does the investor any good until it is realized - that is, until it is converted into Cash. Since that is the case, MattMod was developed to force StdAim to take Cash into consideration within its formulas. To that end, MattMod changes the SAFE value of StdAim from 10% of the Stock Value to 5% of the entire Portfolio Value (Stock Value + Cash).

The result is two-fold: 1) As stock price goes up selling is slowed over that of StdAim allowing more stock to remain active for the investor and 2) as stock price goes down MattMod is slower in getting started but then buying is accelerated over that of StdAim causing both an initial better price of stock and then accelerating the buying of more stock. This seems to correlate well with the way the stock market has existed of these years - that is, the rallies are always greater than the declines. So, MattMod delays selling over that of StdAim to take the larger rallies into consideration and accelerates buying over that of StdAim to take into consideration the shorter rallies.


MACRO AIM by Don Carlson:

MACRO AIM uses the Std AIM program, but only feeds the AIM algorithm when the user-defined Fast Exponential Moving Average "crosses" the user-defined Slow Exponential Moving Average. Thus, the name for Moving Average Cross-Overs (MACRO AIM). Thank you Conrad for the acronym suggestion.

http://www.ttrader.com/mycharts/display.php?p=5462&u=doncarlson&a=Don's%20Charts&id=219


Summary:

Please keep in mind that this study feeds the same parameters to all 100 NDX stocks and only the ending avg PV is posted. I believe that each stock has it's own personality and a unique AIM method that works best with it.

One of the reasons I started this study was to see if the basic Lichello settings could be made more efficient. So far, it appears that the Std AIM settings have to be set to extreme levels before the avg ending PV can beat B/H. The MattMod method is more robust and the avg ending PV can beat B/H by merely setting the Safe to 10%. I'm somewhat surprised that MattMod didn't have a better showing. Most of the comparisons I've performed between Std AIM and MattMod using daily data, MattMod generally exceeded Std AIM by 25%. MACRO AIM exceeds B/H with all parameter settings, and doubles B/H using Std AIM with 10% Safes and updating the AIM algorithm when the current price crossing a 6 month moving average.


LINKS TO CHARTS:

Std AIM Study Summary
http://www.ttrader.com/mycharts/display.php?p=5449&u=doncarlson&a=Don's%20Charts&id=219

MattMod Study Summary
http://www.ttrader.com/mycharts/display.php?p=5450&u=doncarlson&a=Don's%20Charts&id=219

MACRO AIM Study Summary
http://www.ttrader.com/mycharts/display.php?p=5451&u=doncarlson&a=Don's%20Charts&id=219



Other Methods of Interest:

Vortex
http://www.investorshub.com/boards/board.asp?board_id=1341

X_Dev
http://www.investorshub.com/boards/board.asp?board_id=1074

Synchrovest
http://www.investorshub.com/boards/board.asp?board_id=966