Swing Trading Strategies & Stock Picks Since 2003

A reader recently asked me to look into the possible Edge or strategy performance results from using the DMI+ and DMI- crossovers (components of the ADX Indicator).  Let’s see the results, describe the strategy, and see how it performed.

The line of thinking goes that wen DMI+ crosses above DMI-, then an uptrend is in place and that should give you an edge when trading akin to a Trend Following System.  For more information, visit StockChart.com’s page on the ADX Indicator.

The strategy expects to capture large swings from a big trend move and enter you in the direction of that trend.  The system I created (simply) for testing takes advantage of both sides of the market, in that it’s always in the market and goes long when DMI+ (Positive Directional Movement) Crosses Over DMI- (Negative Directional Movement) and then exits and reverses short when DMI+ Crosses Under DMI-.

No stops were used and no commissions were factored into the testing.  Testing runs from January 1998 to present.  The results you see are pure data from TradeStation.

Before looking at the results of randomly selected stocks (by me), let’s look at the strengths and weaknesses of this system on a chart:

(Click for larger image)

The system goes long when the Green line crosses above the Red line (indicator).  It’s expected to capture large swings or trend moves and indeed it does.  This is the DIA around 2004 Daily and it captured two big wins.

Unfortunately, the indicator crossed frequently from the period of June 2005 to September 2005, resulting in numerous whipsaws for small losses.  Are the profitable trades enough to overcome all the small, whipsaw losses?

Unfortunately, no.

I tested this across 8 randomly selected stocks including the DIA and SPY and the results are presented in the following table:

Again, here are the parameters:

1998 - 2009 daily bars; 1,000 shares per trade; no commissions.  Always in the market; no stops.

Over the 10-year period, 3 of the 8 stocks returned positive results, though that would have been eroded once commissions and slippage were factored in.

The Profit Factor… which is the Average Winner (Dollar Terms) divided by the Average Loser (Dollar Terms) is attractive, and in most cases, the Average Winner is at least two times larger than the average loser (giving us an average Reward/Risk ratio of about 2 to 1).

The problem with this strategy lies in the numerous small losses that erode the edge of the larger average winner to the smaller average loser.

The win-rate for these stocks all was less than 33%, meaning only 1 in 3 trades resulted in a profit.  A 2 to 1 reward/risk ratio cannot overcome a strategy with a win-rate of 33%.

Remember, these are just raw data, and if you’re interested, you can run  your own tests and try to include filters such as demanding the ADX must be over a certain value (to filter out choppy environments) or some other method to try to reduce the numerous whipsaws.

Corey Rosenbloom
Afraid to Trade.com