Monday, July 03, 2006

5-Day Up

I like filters to be extremely simple. I also like a filter that returns a managable number of opportunities. For example I could test against the entire spectrum of stocks but I already know that some I simply won't buy because they are either too cheap or too expensive. Too cheap do not move enough in a short enough time to make them worth my time. Too expensive hold a higher order of risk than I'm willing to accept.

Another of my preferences is for stocks that are above their 200 day moving averages.

And I like my filters to follow a specific theme. My method of stock selection relies on the concept of regression to the mean. In other words the market see-saws up and down and in the process always overshoots its null point which is its perfect balance between buyers and sellers. My filters reflect this theme. My very first and as it turned out extremely profitable filter was a simple one - close above 200 day average, close below 20 day average, close above 5 day average.

I also like using a short time period, 5, 10 or 20 days. And I keep the dollar value of the filter at a level where I can buy at least 500 to 1000 shares of any stock I decide to buy using my trading account. And I like the volume to be at least 500000 shares a day.

The 5-day up is a simple filter that looks for stocks where the 5-day moving average is increasing while the 5-day moving average of the 5-day moving average is decreasing.

I tested this filter using the three time standards noted earlier and here are the results.

Direction Start End # Trades Win % ROI % SPX ROI %
Current 3/2/06 06/30/06 699 58 -23 -4.19
Up 3/9/06 5/12/06 443 67 25 5.28
Down 4/27/06 6/30/06 300 51 -67 -14.46

You can see that the win percentage in the up market and the negative ROI in the down market are both significant. You can also see how even a simple filter such as this one can fail during the 6-month period we've just been through.

No comments: