InvestorsHub Logo
Boards
News
Market Data
Markets
Discover
Discover
Subscribe Login/Register
S&P 500
5,303.27
(
0.12%
)
US TECH 100
18,062.00
(
-0.10%
)
Dow Jones
40,003.59
(
0.34%
)
Brent Oil
83.64
(
-0.15%
)
Bitcoin
66,247.61
(
-1.04%
)
Post# of 1453
Next 10
Reply Private New
Public Reply Private Reply New Post
Keep Last Read
Replies (3) Keep Next 10 Report Keyboard Shortcuts
Replies (3) Next 10 Prev Next

aptus

Send PM Follow Ignore
Followers 18
Posts 877
Boards Moderated 2
Alias Born 02/16/2002

aptus

Re: None

Monday, 10/07/2002 1:01:10 PM

Monday, October 07, 2002 1:01:10 PM

Post# of 1453
I've been testing an enhancement that seems to significantly increase returns while reducing the number of trades.

A SIMPLE VOLUME FILTER
It's a simple volume filter that takes a 10 period simple moving average (SMA) of the volume and checks the current volume against this SMA. If the current volume is less, then AIM is not updated. If the current volume is equal to or greater, then we update AIM.

It appears to work well with standard AIM, however the biggest benefit comes when you combine it with Don Carlson's MACRO AIM. Both of these techniques are filters that either allow AIM to be updated with the current price (and volume), or block the update (as if the update was never done).

Adding the Volume filter to MACRO AIM means that there are now two filters that must be satisfied before an update can proceed to AIM. If either filter "catches" the update, then AIM never sees the update request and an AIM calculation is not performed.

WHY ARE FILTERS GOOD?
Although AIM is said to be "Automatic," it really isn't. In fact there are a number of important, and subjective, decisions that must be made when setting up a new AIM portfolio. The two most important are:

1. Equity selection.
2. Model selection.

Once the portfolio setup is complete there is the question of how often to update the portfolio. Generally it's better to update more often, however doing this has two major disadvantages.

1. You can run out of cash in a long-trending down market.
2. AIM will most likely recommend some trades that don't add to your returns and reduce your risk (they might even be detrimental). In other words they're inefficient trades.

Filters get around both of these problems. When using a filter, you're free to update as often as you want -- I recommend daily updates. The filter takes care of when to invoke the actual AIM update. This can lead to a more efficient trading pattern (depending on the filter, of course).

WHY A VOLUME FILTER?
As I've mentioned in the past, I believe AIM is missing a significant amount of information by just looking at price when making its decisions.

Price is important, but doesn't tell the entire story. For example, a price increase of 50% on a volume of, say, 50 shares has a high degree of standard error associated with it. Whereas a price increase of 50% on a volume of 50,000,000 shares has a relatively low degree of standard error.

The reason for this is because, with the 50 share example, only a very small number of buyers (let's say two buyers) have said they think the price increase of 50% is justified. Perhaps everyone else disagrees and won't buy the stock at its new price level. So if you owned these shares, there is a low probability that you would find a buyer for your shares at the increased price.

With the 50,000,000 share example, a relatively large number of buyers have said the price increase is justified (e.g. 1 million buyers). This means that there are significantly more people out there willing to purchase shares at that price. Thus you're more likely to find a buyer if you decide to subsequently sell at the increased price.

And that is why volume is an important piece of information. If you're trading only on price, you're missing a SIGNIFICANT piece of information, information that's available to you free of charge.

WHY USE MULTIPLE FILTERS?
Let's put everything together and see where we end up.

First, we have AIM. AIM is an excellent algorithm for taking advantage of volatility and signaling when to trade. However it requires a number of subjective inputs (as we've seen above) and in its standard form has a difficult time beating the Buy and Hold strategy for all but the most volatile stocks (before anyone jumps all over this statement, run some historical tests on the DOW, NASDAQ and S&P500 stocks using standard Lichello AIM and view the results). Therefore standard AIM will work well only when we choose certain types of stocks -- and that means we're limited to a small portion of the available stocks.

Secondly, we have the MACRO filter. This filter decides the best time to update based on a pair of Exponential Moving Averages (EMAs). When a significant event occurs in the MACRO filter, AIM is updated, otherwise it's not. This has the advantage of only updating AIM when the current price determines it's most efficient to do the update. Inefficient trades are thus filtered out.

And finally we have the Volume Filter. This filter decides the best time to update based on volume. The advantage is that price moves that have a high probability of being inefficient are filtered out.

Therefore we are filtering based, not only on price but, on price and volume. Each filter is only letting efficient trades through to AIM. AIM, then, acts as a final filter to determine whether a trade recommendation is given.

CONCLUSION
So, if the theory is correct, adding these two filters to standard AIM should provide significantly more efficient recommendations (i.e. we're filtering the inefficient trades) and thus increase our returns and lower our risk. Does the experimental evidence support this? Let's find out by looking at a number of tests.

1. A 5+ year test of the DOW.
AUTOMATIC INVESTOR ANALYSIS RESULTS

Total Portfolio Value: $431,078.11
Total Investment: $300,000.00
Simple Return: 43.69%
Buy and Hold Total Value: $370,015.55

2. A 7+ year test of the DOW.
AUTOMATIC INVESTOR ANALYSIS RESULTS

Total Portfolio Value: $839,515.67
Total Investment: $300,000.00
Simple Return: 179.84%
Buy and Hold Total Value: $708,086.89

3. A 10+ year test of the DOW.
AUTOMATIC INVESTOR ANALYSIS RESULTS

Total Portfolio Value: $1,104,573.05
Total Investment: $300,000.00
Simple Return: 268.19%
Buy and Hold Total Value: $970,957.39

4. A 5+ year test of the Nasdaq 100.
AUTOMATIC INVESTOR ANALYSIS RESULTS

Total Portfolio Value: $4,371,797.05
Total Investment: $970,000.00
Simple Return: 350.70%
Buy and Hold Total Value: $1,658,227.63

So far the filters are looking good.

THE FUTURE...
The next step is to test these filters over different time periods (e.g. 1969 to 1979). I've run some tests over various periods but haven't come up with a definite conclusion. However I can say this, in all tests to date, adding the filters to AIM has produced SIGNIFICANTLY better results that just using standard AIM alone.

Regards,
Mark

http://www.automaticinvestor.com
Public Reply Private Reply New Post
Keep Last Read
Replies (3) Keep Last Read Next 10 Report Keyboard Shortcuts
Replies (3) Next 10 Prev Next

Join the InvestorsHub Community

Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.

Sign Up