The other portfolio manager, the “greater light,” was Elaine Garzarelli. She said that it was no longer a viable strategy to simply buy and hold any stock. Every stock has to be monitored, and is an eligible sell candidate if it “misbehaves” (begins to “break down”). The last market crash should be sufficient evidence of the shortcomings of the former strategy and of the correctness of Garzarelli’s position. In the crash, we saw perfectly good companies decline in value 60% and more. We have also seen volatility soar to unprecedented levels. Under the circumstances, what justification is there for going forward with the attitude that most positions should be held for several years? Typically stocks do not go up non-stop for several years. Over that period of time most stocks will have several serious corrections in which a large portion of previous gains will evaporate.
It makes much better sense to lock in gains when you have them if a stock becomes unusually weak and switch to something else rather than to become a stock market yo-yo (the new “sell on weakness” school of thought). Also, since no one can be sure how far a stock will drop once it begins a decline, the well-considered positioning of protective stop-losses is in order. If the stop-loss is ratcheted up as the stock rises, an increasing amount of the gains will be “locked in” should the stock decline.
The author once spent more than eight hours a day for three years testing tens of thousands of computer-driven investment strategies. The goal was to develop systems that worked well regardless of the state of the market or the general direction of the stock. The systems also had to outperform a “buy and hold” strategy by a wide enough margin to justify their use. The goal was to find strategies that could outperform a “buy and hold” approach by at least 20% a year. Thousands of tests were performed on each of thousands of stocks over a wide variety of market environments covering a period of many years to get a good profile of each strategy.
The more successful of these automated systems had something in common. Each had many losing trades, sometimes far more losing trades than winning trades. This was a surprise. Many gains were expected, not many losses. Their trading patterns over many years revealed several other interesting characteristics. First of all, the systems with the best results were “Nervous Nellies.” That is, they sold at the slightest provocation (usually a downward motion of the stock that satisfied certain pre-established criteria). The fact that there is so much “noise” or non-significant motion in stock behavior is what generated a large number of the small losses. Yet, some of these same systems could generate large annual gains in a severely declining market.