Part I: In the Beginning

Table 3.1 Possible Time Frame Combinations

way the terms really referred to the types of bars depicted on the charts, not to the charts themselves. Thus a short-term chart used daily bars by definition. In the early eighties, the pace of change started quickening. The demarcation was the introduction of stock index futures trading, with the birth of the ValueLine futures traded on the Kansas City Board of Trade. Short term started meaning hourly charts, intermediate term daily charts, and long term weekly or monthly charts. In the intervening years the trend has continued relentlessly toward ever-shorter time frames. See Table 3.1 for possible time frame combinations. However, no matter what the time frame, the underlying concepts are approximately the same.

For example, long term is the time frame in which you do your background analysis. It is the environment in which you determine your overall outlook and the broad strokes of your investment plan. For investors with long horizons, monetary and fiscal policy figure importantly, as does the flow of funds, valuation data, and the regulatory environment. For investors with a shorter horizon, important factors might be the direction of the 200-day average or the slope of the yield curve.

Intermediate term is the time frame in which you do your security analysis. It is the time frame for stock selection and group rotation. Broad market statistics can be important here. Long-horizon investors will consider broad market data such as advances and declines, new highs and lows, sector rotation, relative-strength trends, and quarterly supply and demand factors. Shorter-term investors may be looking at consolidations, turning points, and breakouts in industry groups.

Short term is the time frame in which you execute your trades. It is the time frame you use when placing your orders and seeing

Chapter 3: Time Frames

to the optimum execution of your strategy. This is usually the province of short-term technical indicators, price patterns, changes in volatility, trading data from the floor, etc.

Each time frame has its tasks, and those tasks, along with the tools used to accomplish them, will vary from investor to investor. What is most important is to keep each time frames tasks separate and distinct. A prime example of breaking this rule is to continue looking at the short-term chart after the trade is executed! After execution, your focus should shift back to your intermediate-term tools, as these are the tools you maintain your trade with. Only when your intermediate-term tools and techniques call for exiting the trade, either to take a profit or to prevent further loss, should you turn back to your short-term tools to execute the decision.

The blurring of the tasks in combination with the mixing of time frames actually makes investing harder. It confuses the decision-making process and clouds thinking. Often when the time to make a critical decision is at hand, the temptation is the strongest to abandon discipline and use a tool or tools in a manner for which they were not intended. While this may seem to add information, the bottom line is less reliable information. The new data acts to muddy the waters with conflicting information not well matched to the task at hand.

From an analytical perspective, these ideas have important impacts. Bollinger Bands can be used in all three time frames. They can be scaled to suit in three ways, by choosing the time period represented by each bar, by choosing the number of bars used in the calculation, and by specifying the width of the bands. The base for Bollinger Bands ought to be a chart with bars coincident with your intermediate time frame, the base time frame for the calculation ought to be the average that is best descriptive of your intermediate-term trend, and the width ought to be a function of the length of the average. In our shop, daily bar charts, a 20-day calculation period, and 2 standard deviation bands are typical.

Note the use of the term descriptive in the paragraph above. Do not try to pick the average that gives the best crossover buy and sell signals. In fact, the average we want is considerably longer than the average picked by an optimizer looking for the greatest profit from crossover signals. Why is this? Because our signals will come from interaction with the bands, not from crossovers. The average we select is used as a base for building a relative

Part I: In the Beginning

framework within which we can evaluate price action in a rigorous manner. This average will be better at defining support and resistance than at providing crossover signals.

The best way to identify the correct average is to look for the average that provides support to reactions, especially the first reaction after a change in trend. Suppose the market makes a low, rallies for 10 days, and then pulls back for 5 days before turning higher again and confirming the birth of the new uptrend by taking out the high for the initial 10-day rally. The correct average would be the one that offered support at the low of the 5-day pullback (Figure 3.1). An average that was too long would have been too slow to define support, and too slow in turning higher to describe the new trend (Figure 3.2). An average that was too short would have been crossed three or more times, and would not have given useful support or trend information (Figure 3.3).

In studies done many years ago, the 20-day average proved to be a good starting point for most things financial. The adaptivity of Bollinger Bands comes primarily from volatility, not from moving-average length selection; so we want an average length

-Data -*-ldeaT

Figure 3.1 Moving average, correct. Price crosses the average shortly after the low and then provides support on the first pullback.