Friday, July 12, 1996

A Relative Strength Update
by MF Dowman

In an earlier Fribble, "Strong Relatives," I outlined a new variation of the classic Investing For Growth approach. As you may know, we're currently working to back-test the Relative Strength variation, but all things point to it being an improvement over the original model. We'll know more historical details in the next few months as soon as I can plow through old databases to test it.

In my earlier Fribble, I speculated about using absolute cut-off points as buy and sell signals and many new readers to the approach are asking questions about those points. I don't think the cut-off points are really necessary and they complicate the approach, so let me take this chance, then, to outline the approach as we're using it in the model portfolios.

If you've read the original IFG Primer (and if you haven't, you can get a copy in our FoolMart), you know the reasons why we use the Value Line Timeliness Rank and the High Growth Stocks screens. None of that has changed for the Relative Strength variation.

Step One is still to take all of the stocks from Value Line's High Growth Stocks screen with a Timeliness Ranking of 1.

Then, instead of the classic approach of screening those stocks by Value Line Industry Ranking, Step Two for the new variation is to rank those stocks from Step One by the Relative Strength score from Investor's Business Daily.

Step Three is to select the top ten stocks on that list (if there's a tie for the final spot, then turn to the Timeliness Ranking). One advantage of this approach, though, is that it allows you to select fewer than ten stocks if your portfolio's growth component doesn't need, or can't sustain costs on ten stocks. If you want five, take the top five.

On the sell side, you can update your portfolio as often as monthly or as infrequently as quarterly. From what I can tell with limited data, updating the portfolio more often increases the performance by rotating out of "mistakes" more quickly and picking up the new strong stocks. Updating more often, however, does increase trading costs, so you'll have to choose an appropriate holding period for your own situation.

The way to determine how often you should update is by measuring your trading costs. If you can keep your trading costs under our benchmark of 2.5% of your capital (annually), then updating monthly may be the best way. If your portfolio is smaller, then push the update cycle out further, say every 6 or 8 weeks, or quarterly to suit your particular situation. There's nothing magical about a particular holding period. The theory works on the idea of rotation into the best stocks at any given time. Updating more frequently (say every month), seems to improve the model's sensitivity and performance, despite the added trading costs (assuming you're paying no more than $20 or so per trade).

Stay tuned and as the back testing is accomplished, we'll keep you posted.

Transmitted: 7/12/96