In Dr. Martin's Zweig's winning on Wall Street, Dr. Zweig presents a trading idea that was inspired by the work of this friend Ned Davis.  It is a very simple trading method based on weekly data from the Value Line arithmetic index.  The rule are quite simple and the results are quite impressive.

Buy - Buy when the weekly changed of the Value Line index is greater than 4% up

Sell - Sell when the weekly change is greater than 4% down

Hence the creative name the 4% model.  Remember this system was written back in 1986.  So how well looking forward did this system do?

 

 

I only have weekly data going back to 1992, but the 500% return is not so bad, and visually the draw-downs look tolerable.  The system is having it biggest problem since 1992 coming into this year as it struggles with the volatility and several 4% weeks causing it to whipsaw, but look at the value of the equity today... the equity curve is at an all time high!  Boy what I would do to be there!  The system bought in on the 3/13 bar this year and has not been shaken out yet since we have not had a 4% weekly retracement yet.

I will begin tracking this model as part of this website so we can follow it's signals.  It very rarely fires a buy or a sell.  On average it stay in a trade for 34 weeks.  In a winning trade it stays in on average 49 weeks, almost an entire year.  For a losing trade it tends to be shaken out in about 6 weeks.  So far this system has been in the current trade for 9 weeks.

To track this system we will use the Value Line arithmetic index whose symbol for TradeStation is $VAY.  We will use $VAY to trigger our system and we will buy RRY which is the Ryder 2x Russell mutual fund.  I also added a 10% stop to the strategy.  Had we traded this system since March this would be our current trade status:

Those of us who use the trin on an intraday basis know about its glitches as large volume stocks move from the advancing column to the declining volume or versa-vice.  On Thursday this happened when Citigroup hovered around the previous day's close.  It caused the trin to shoot across the magical 1.00 level from .86 to over 1.20.  Since a stock can be in only one of two states, either advancing or declining, a "state change" causes the entire volume to shift.  For most stocks this has little effect on the trin value, but with a high volume stock like Citi it can have a profound effect causing a quantum jump in the indicator.

We have identified the first problem with the trin but there is another issue with this indicator, volume.  Since volume is not dollar weighted in the trin a share of Citi at $4.00 has the same impact as a share of Goldman Sachs at $140 per share. This can swamp the indicator with low dollar high volume stocks thus not accurately present the internal machinations of the markets.

The low price of the financial's and the high volume from their daily churning have led to record lows on the trin as we have moved through this bull market and I been reluctant to use any level information only intraday trend information for trading.

 

As a project I will add to my list producing a better trin indicator.   Currently, for the record, I do apply an inverse function, (ie. I multiply the trin by -1), this allows me to see the trin as a having a direct relationship to prices.

-rlt