102 BUILDING YOUR KNOWLEDGE FROM THE GROUND UP

"The daily net A/D - Advancing issues minus decline issues on the NYSE.

The net A/D from lO days ago.

•The current 10-day moving sum. Todays A/D, minus the net A/D from ten days ago, plus yesterdays moving 10-day sum. Example: For 7/24, -2 - (-440) -(- (-1556) = -1118.

The 30-day sum of advancing issues minus declining issues as of 30 days prior to the current date.

The current 30-day moving sum. Todays A/D, minus the moving sum 30 days ago, plus yesterdays moving sum.

The 10-day equivalent moving sum. Todays 30-day moving sum divided by 3.

"The difference between yesterdays close and the close from the previous 5th day on the New York Composite Index,

The lO-day moving sum of the values in column A.

The 10-day net moving sum. The difference between todays value of the lO-day moving sum, and the moving sum from ten days ago. For example, on 8/3: 61866 - 63937 = -1531.

For the longer term, I do the same thing for a 30-day period and divide the result by 3 to obtain a 10-day equivalent average, as shown in the Table 8.2. Plotting the results gives a chart like that shown in Figure 8.13.

A quick comparison of the breadth momentum lines to the daily closing prices of the Industrials shows a good correlation between tums in the oscillators and intermediate price r eversals on the Dow. You will note, however, that if you traded on the basis of tums in the short-term oscillator, you would get "whipped out" on several occasions. The long-term oscillator, on the other hand, sometimes lags changes in intermediate trend, such that trading by it alone would cost you money. Thats why I use them as supporting tools. But overall, the correlation is very good and has been since I first began using it in 1975.5

The price oscillator I use is slightly more complex, and it is als more important in terms of the weight I place on it in evaluating the market. Again every moming, using the daily closing prices on the New York Composite Index, I begin by calculating the difference between the previous days closing price and the closing price from the previous fifth day. Then, I add the result to the sum of the same result for the previous nine trading days, to give a 10-day mnning sum of the five-day price difference (see Table 8.2). A plot of the daily result compared to the Indu strials average shows a highly significant correlation between tums in the oscillator and tums in the daily closing prices of the stock index (see Figure 8.13).

I use the breadth and price oscillators to anticipate coming market tuming points and to con firm them when they occur. A simple look at the charts will show that the higher or lower the oscillator goes on the chart, the greater is the chance of an intermediate change of trend. Like a waming light on railroad tracks, when the oscillators approach or break through significant previous highs or lows, they are telling you that danger is ahead. They dont tell the nature of the danger, or how far ahead it lies, but they do tell you to proceed with caution or continue full speed ahead.

For commodities, as I have already mentioned, my quote system has an oscillator measurement built-in to it which is the difference between two moving averages. The nice thing about this feature is that it allows you to readily change and experiment with different time periods for the moving averages until you find one that fits the movement of the specific market you are dealing with.

This is essential in the commodity markets because, any time you are dealing with one item versus an average, the chances of conditions arising which change the nature of "normal" price behavior are much greater. Oscillators in the commod ity markets are, therefore, much less reliable than they are for the stock average indices. Nevertheless, if you find the right one, they can be of tremendo us value in confirming trend changes.

In fact, sometimes you can find a correlation that is so good that you could virtually make your trading decisions on them alone and make quite a bit of money. I never do this, nor do I recommend it; but it is possible, and Im sure there are

many traders who do it. The problem with this approach is that when you are wrong, you will often be wrong in a very big way.

Many technicians get very sophisticated with oscillators and use them, not just for the averages, b ut for specific stocks as well. I dont.

There is such a thing as acquiring too much information. I think you are better off using a few oscillators at most, and then only as secondary measures to affirm or deny the primary change of trend indicators I discussed in the last chapter. Frankie Joe had a trading mle which he called KISS, an acronym for "Keep it simple, stupid!" Good advice from a man who was what I like to call "a pros pro."

MAKING SPECIFIC STOCK SELECTIONS

So far, all of the methods Ive discussed apply to virtually any market, from specific stocks, to commodities, to indexes. Undoubtedly, some of you trade only individual stocks -something I used to put my primary focus on. For making stock selections, so far we have Dow Theory to make a ge neral market call and, within that context, the technical methods Ive described to pick out a group of stocks that are likely to move with or against the market. Now, well cover a few auxiliary methods which will

help you select specific stocks. The objective, of course, is to put even more of the odds in your favor. The Technical Versus the Fundamental Approach

There is a whole group of traders, some independent, some working for firms, called stockpickers-people who choose stocks for trading, speculating, or investing. Within this group, there are two basic schools of thought-the purely technical school and the purely fundamental school. My experience has been that it is a very rare purist who consistently makes money.

Most profitable stock speculators are hybrids; they combine the best of both worlds and use a combination of technical and fundamental tools. I would have to call myself a hybrid who leans toward the technical side. I combine the technical methods described in the last chapter, which are likely to reflect the future judgment of market participants, with fundamental statistics which consistently correlate to price movements over time.

The fundamentalists believe that, over the long term, the market prices stocks according to yield, eaming power, and the value of each companys underlying assets. In other words, value is determined by those three intrinsic factors for any stock. The problem I have with this approach is that it totally ignores the subjective nature of value; it doesnt take into account that people, not computers, ultimately determine price. For the fundamentalist view to be accurate, yield, eaming power, and assets would have to directly reflect the collective judgment of the market

place. They dont. In fact, fundamentalists are notorious for being "right." but with bad timing.

One very popular fundamentalist offshoot is the Graham and Dodd approach to stock selection. Highly oversimplified, the Graham and Dodd approach says that you should buy low PE (price/eamings ratio), low book value stocks. The underlying premise is that these are likely to be the "undervalued" growth stocks, whereas the higher PE stocks have already been recognized and properly "evaluated" by the market. As long as you are in the early to late-middle stages of an inflationary bull market, this approach works pretty well stock prices for all but the worst of companies appreciate. But over time, it can get you in trouble. Usually, when a stock has a low PE and a low book value, there are reasons for it, and the market already knows those reasons.

I find PEs and book value very useful but in a different way. If you look at PEs and book value of the indexes as a whole and compare them to historical PEs and book values, you find a good secondary indicator of overbought and oversold conditions of the averages as a whole. Then, you can compare the individual stocks PE and book value to those of the averages, and gauge, again as a secondary (perhaps tertiary) indicator, the relative performance of the specific stock to the market in terms of an overbought or oversold condition.

Rate of Change of Eamings Growth

One fundamental statistic that is excellent is the correlation between rate of change of eamings growth and the change in the stock price. In a book published in 1969, Gordon Holmes demonstrated that, "The slope of a given price trend almost always precedes the correspondent or equivalent eamings trend slope in time. The amount of time displacement is about three months."6

There are three things I like about this observation. First, it is usually true. Second, it has withstood the test of time. And third, it supports the fundamental precept of Dow Theory that the markets "discount everything." Now, how do you use this observation?

First of all, you have to establish that the correlation holds true for the stock you are evaluating. Most businesses have seasonal fluctuations which cause variations in eamings, so you should look at eamings figures for no less than six quarters before considering eam ings growth as a valid indicator of future price changes. A plot of the eamings curve (lagged three months) on the same time base should give you a direct correlation between eamings growth (or decline) and price growth (or decline). The correlation should exist for the entire period considered, otherwise this indicator should be disregarded completely for that stock.

Holmes developed a rather complicated method for stock picks based on eamings growth and other factors; I use eamings growth a little bit differently. In William ONeils New York Stock Exchange