Terminology

For those who are first time reading Trend Watch, please click here to understand the term definition and/or how to interpret it.

Part I - Market Comments

It was another difficult week for the stock market. The action was very poor in the first half of the week. On Thursday the indices stabilized and on Friday the market rebounded significantly. For the week, the Dow Jones Industrial Average lost about 5 points, basically remaining flat. The S&P 500 lost 0.8%. The Nasdaq Composite finished 1.2% lower. The Russell 2000 (the small capital index) was the biggest loser last week, ending 1.7% lower. By far, the year-to-date performance of the Nasdaq Composite is -5.1%. The US stock market is really mean this year.

Turning to WATTS, what did the system say about last week? It's simple and clear -- the signal readings went down stably for both short- and intermediate-term systems, even on Thursday and Friday when the market rallied. Usually, if the market rally were just a short-lived rebound in a continual downtrend, the signal readings of WATTS would just budge a little and wouldn't raise any red flag to us. So far, that is the case with the system.

The major part of the WATTS system is the trend timing model used to follow the short-term and longer-term market trends. Since it's a trend timing model, it is only able to follow the trends that market demonstrates. In other words, if you ask when this current market decline will end, you couldn't get answers from this part of the system. It is not the responsibility of the trend timing model to tell us what will happen. It can only tell us what is happening.

Coming out of greed, one of the very common human natures, we tried to push the system performance and make it more profitable when we were developing the system. We developed another model to detect the potential short-, intermediate-, and long-term bottoms for the US stock market in addition to the trend timing system. We then incorporated this model into WATTS and called the signals generated by this model the Bottoming Zone Alert.

The reason to develop and incorporate such a model into WATTS is from an interesting observation to the market. The stock markets usually take longer time to develop and form the top. Conversely, when the stock markets reach the bottoms, the following rally or rebound is usually very sharp and violent. A trend following system loses a somewhat significant partial profit from its bear signals. To void it, the system must be able to indicate the potential bottoms sooner.

The Bottoming Zone Alert is acting on a real-time basis. That is, when the market reaches its bottom, the system issues the Alert at the same time. There are 3 types of market bottoms in terms of time scale - short-term, intermediate-term, and long-term. Therefore, the system can issue the different levels of Alerts to these 3 bottoms, respectively. (Please read the terminology in the beginning of the newsletter to get more details about it.)

After our back-testing and some live alerts, we found that our model can literally pick up the market bottom to the day with one or two days away at most.

There is, however, no free lunch and perfect stock investing system in the world. Well, some may argue that there surely are free lunches. But we are pretty sure of that there is no perfect investing system. There is a caveat coming with the Bottom Zone Alert. After WATTS issues the Level I Alert for the potential short-term market bottom, it doesn't mean that the market will stop falling. From there, the market can keep falling and thereafter make the system issue its Level II Alert for the potential intermediate-term market bottom, and so on so forth. Precisely, a Level I Alert means that if the market decline decided to stop, that's where it stopped and you could enjoy a strong rebound from there.

Besides, not all the short-term market bottoms will be detected by our model. Most of the minor market bottoms will be skipped on the radar screen of the Bottoming Zone Alert model. For example, this model only issued 2 Level I Alerts this year by far, one was in January and another was in April. Those two were considered as significant short-term bottoms. The others like the rebounding bottoms in the downtrend since August were just not strong enough for our model to pick them up.

Why are we talking so much about this model? We have a good reason. After the last issued Bottoming Zone Level I Alert in April, we are now approaching to where the Level I Alert will be possibly issued in the near future. The signal reading is still a little far away from where it would issue the Alert, but it's certainly in the neighborhood where we would like to pay full attention.

Nothing is guaranteed in the stock market. But if the market continues to decline next week, there is a good possibility of that WATTS will issue its third Level I Alert this year. We don't know fore sure that if it will happen or even though it does happen, we don't know if that's final alert for this current decline. The market can, of course, go on to drop more to make WATTS give birth to the Level II Alert, which is quite rare.

There will be more clues being released next week for us to judge the future market direction for the rest of the year 2005. Last year, we had a surprising rally into the end of the year. What may we have this year? Judging from the conclusion of that the current downtrend would be a long and strong one and by far, the current SELL signals only yield about 4% return, we incline to believe there is more downside to go.

Anyway, the SELL signals of WATTS are solid and locked, please hold your positions tight. Let's just follow the trend.

Part II - Signal Update

WATTS Short-term Signal Status

Last Closed Short-term Signal

WATTS Intermediate-term Signal Status

Last Closed Intermediate-term Signal

WATTS Bottoming Zone Alert

Part III - Model Portfolio Update

Model Portfolio provides investor who doesn't know how to react to the WATTS's signal a simple guideline of how to manage his/her portfolio. If you are interested in following this guideline table, please Read This Important Note first.

Model Signal	Account Type	Equity Type	Most Aggressive Portfolio	Least Aggressive Portfolio
WATTS Short-term Signal	Stock Trading Account (Margin Account)	Stock & ETF	50% QQQQ short, 50% cash	0% SPY, 100% Cash
		Mutual Fund	None	None
	Stock Trading Account (Non Margin Account) Self-managed Retirement Accounts (IRAs, 403b...)	Stock & ETF	0% QQQQ long, 50% Cash	0% SPY, 100% Cash
		Mutual Fund	50% POTSX	None
	Company-managed Retirement Account (401k)	Stock & ETF	Not Available	Not Available
		Mutual Fund	50% POTSX, 0% High-growth Fund, 50% Money Market or Short-term Bond Fund	0% High-growth Fund, 0% Balanced Fund, 100% Money Market or Short-term Bond Fund
WATTS Intermediate-term Signal	Stock Trading Account (Margin Account)	Stock & ETF	0% QQQQ, 100% Cash	0% QQQQ, 0% SPY, 100% Cash
		Mutual Fund	None	None
	Stock Trading Account (Non Margin Account) Self-managed Retirement Accounts (IRAs, 403b...)	Stock & ETF	0% QQQQ, 100% Cash	0% QQQQ, 0% SPY, 100% Cash
		Mutual Fund	None	None
	Company-managed Retirement Account (401k)	Stock & ETF	Not Available	Not Available
		Mutual Fund	0% High-growth US Stock Fund, 100% Short-term Bond Fund or Money Market	0% High-growth US Stock Fund, 100% Short-term Bond Fund or Money Market

PART IV - Optional Reading Material

1. Leading or following?: http://www.marketwatch.com/news/story.asp?dist=&param=archive&siteid=mktw&guid=%7B78DB13F1%2DAF34%2D49A2%2D928E%2DC9ED511473CB%7D&garden=&minisite=

This is a very excellent market commentary on a long-lived trend timing system and how to recognize its modern value in a fast-changing financial world today. On Monday, this well-known trend timing system issued its sell signal. Should we pay attention to its sell signal? From the track record of the trend timing system, it said so.

Looking at the entire track record of more than 25 years that the Hulbert Financial Digest has for the trend timing system, its performance for the US equity market is in first place for risk-adjusted performance among all timing systems the HFD has tracked. Simply speaking, it's number one over the more than 25 years of period as a trend timing system.

But the commentary continued to argue that such a point of view is quite deceiving. The main reason? It said that the outstanding past performance of the system was mostly from its first 15 years or so. The last 10 years was so-so, to say the best.

This article forces us to ask: when would an outstanding trend timing system fail to deliver its past glory? Nobody knows the answer exactly. But we believe that the success of a system will destroy itself if there are too many people following the same successful system. That's why we keep our proprietary model extremely confidential. In the future, we may even limit the number of our subscribers to avoid this fatal problem.

Besides, we provide a simple explanation to the shabby performance of the above system in recent years. Evidently from its past year-to-year performance data in the article, the system worked great in a primary bull market. It started working poorly since the late period of the bull market. Its performance in the primary bear market since 2000 is also quite poor. Usually, this explains that the above system is a bull-market timing system. It works very well in bull markets. But in the late period of a bull market when the market volatility significantly increases and the big money is moving out of the equity market slowly and quietly, it loses its ground to be a good timing system.

That is a very common problem to the market timing systems. Bull market, bear market, and the flat market between the prior two markets have quite different characteristics internally. A successful bull timing system could be a disaster in bear or flat markets. If the developer of a market timing system doesn't take good care of this issue, the resulted system will yield an unstable performance in the long term. We did keep this in our mind when we developed WATTS and obviously we got some kind of handle on this problem.

This article also partially answers to an often-seen question of the historical back-testing of the mechanical system. Does the back-testing mean anything? Can it be trusted? Well, let's keep this interesting argument for the future issues.