Greetings, fellow Pro-fundity team members -
4-30-99 Page

Technical Analysis - Trading bands, Bollinger bands, Dow theory

Review:

1. Last week we discussed using averages to understand market trends. The trend in a price chart tells us the direction the price is moving. This is helpful as we consider the purchase or sale of a stock.
2. There will be times when a stock price (or a total market) will move sideways, without any apparent trend. These are periods of indecision, often merely a pause in the existing trend. Other times the pause may forecast a change in direction. The figure below, of a recent pick, shows such a price pattern:
   
3. What looked like a good up-trend faltered and seems to be waiting for some signal. This stock was picked because of this pattern. These periods of indecision provide great rolling opportunities. However, the indecision is not permanent. We may only get a few more rolls.
4. To help spot these periods of opportunity, we introduced the Moving Average last week. This tool in our bag of tricks helps identify trends. However, even before the moving average, we can spot trends by eye and draw them manually. The next figure shows the price chart we used last week with eye-balled trend lines:
   
5. Notice how the trend lines are drawn on the price chart. An up-trend is connected to the bottom of the price pattern, while they are connected to the tops on the down-trend. The reason for this convention is we are looking for a break in the trend as a trigger. On an up-trend, the change will be a break down through the trend line. The reverse is true on a down-trend.
6. The moving average helped identify these triggers with less eye-ball and more data. We repeat the 15 day MA chart below for NSPK. Triggers occur where the moving average crosses the price line. When the price line moves up through the MA, this signals bullish strength and is generally a buy signal. The opposite occurs as the price moves down through the MA.
   
7. On most charting software we are given the option to select Simple or Weighted moving averages. The simple MA was the focus of last weeks Guidepost. The difference between these two is one of choice where some analysts prefer to give more weight to recent price action. The "exponentially smoothed" is the most common weighted average which we will feature in our May 21st editorial. At this point just understand that it is available. We can experiment with these factors as we develop our own trading strategies.
8. The moving average plays another role in generating trading signals, using two MA's together. The strength and/or weakness of a trend can be implied from the relationship between the two MA's. Investors commonly use 200-day and 50-day MA's together and study their relationship. The shorter MA is more sensitive to change, following the price chart more closely. The longer MA smoothes the pattern more. When these two MA's cross on a price chart a signal is generated. For instance, when the shorter MA crosses above the longer, that signals strength and is bullish.
9. Traders, those dealing with shorter time frames, use the same principle but with shorter MA's. 7-day and 15-day or 10 and 40-day pairs are common. The price chart for NSPK illustrates the principle. Look carefully at the crossover points of the two MA's. The concept seems to work, however the signal is always late. This has been the case in each example of the moving average which is a lagging indicator. We need to improve this strategy to make money.
   
   Trading Bands
10. A trading band, or envelope, is created by adding parallel moving averages above and below the center MA. The envelope then defines the upper and lower limits of the normal trading range for a stock. The width of the envelope can be adjusted to account for the volatility of the price pattern. When the stock price reaches the upper band a sell signal is generated. Likewise, as the price decreases to the lower band, "buy" is the signal.
    
11. Charles Dow outlined the reason why trading bands work in a series of articles in the Wall Street Journal back at the turn of the century. The Dow theory was built around several assumptions regarding the general stock market that work equally well on individual stocks. The most important for this discussion are:
    1. The market is comprised of three trends, Primary, Secondary and Minor. The first usually lasts more than a year and as long as several years. A bullish primary trend will have successively higher highs and higher lows. A bearish primary trend will have the opposite. Secondary trends are one to three month reactions within the primary, where Minor trends last from one day to three weeks. Dow suggests the minor trends are unimportant and often misleading.
    2. Primary trends go through three phases, a result of market psychology. First, price is down, gloomy outlook, lack of enthusiasm. Smart investors load up on the bargains. Second phase reflects increased corporate earnings and good economic conditions, continued accumulation by savvy investors. Third phase finds even better earnings, improved economy and now an encouraged public who see the price going higher and higher. This leads to a buying frenzy while the savvy investor begins to unload expecting the prices to drop.
    3. Trends remain in place until acted upon by a significant market force. This is termed momentum, or the tendency of a moving body to remain in motion unless acted upon by some outside force. Longer trends are more likely to continue in their direction than shorter trends. Trends will continue until a definite signal for reversal is given.
12. We see from the chart above how price seems to bounce off the trading bands. The trading envelope becomes support on the bottom and resistance on the top. Support and resistance lines represent a resistance to change. As the trading envelope follows the general price trend, we see how the principle of a rolling stock is not limited to a trading band where support and resistance remain level.
13. There are periods in the chart, however, where the price is less volatile and does not come close to the bands. That is, where the price pattern is not changing much, the trading envelope is of little value. A trading envelope with constant band width leaves some of the price pattern out of the picture. John Bollinger created a trading envelope where the width changes with volatility, called Bollinger Bands.
14. The difference between Bollinger Bands and envelope trading bands is that envelopes are spaced some fixed percentage above and below a moving average, whereas Bollinger's are plotted some number of standard deviations above and below the MA. Since standard deviation measures variability, the bands adjust themselves, widening during volatile times and narrowing during calmer periods.
15. Bollinger uses the standard deviation of the price data to set the bandwidth of the envelope. See the sidebar at the end of this guidepost for an explanation of "standard deviation." This makes the trading envelope more sensitive to price fluctuations.
    
16. The basic principle is that prices tend to stay within the upper and lower band. The spacing between the bands varies with the volatility of the prices. During periods of extreme price changes (i.e., high volatility), the bands self-adjust, widening to contain prices. During calmer periods, these same bands contract to further contain prices.
17. With the Bollinger Bands, we can select the width of the band as required to optimize the trading signals. Bollinger recommends a band-width of 2 standard deviations (that is 2 above and 2 below) and a period for the moving average of 20. The chart below shows Bollinger Bands on another recent pick, SUPC.
    
18. Notice on this chart how the relationship between price changes relative to the trading bands provide information about where the price is likely to go. It is this relationship that provides value in forecasting price changes.
19. To use trading bands on TC-2000, follow this example:
    1. First select the ticker SUPC by clicking on the heavy J icon (for Jump to ticker), typing SUPC, then OK.
    2. Click on the heavy I icon (for Indicators).
    3. Click on the line Prices - Bar Chart.
    4. Click on Add Indicator.
    5. Click on Envelope Channel, then OK.
    6. You now have a new line under Prices - Bar Chart
    7. Select the setting "15" under Period and "20" under Width.
    8. Choose a light drawing color so the trading bands can be seen against the dark window background.
    9. Be sure there is a check mark in the Visible window.
    10. Choose Simple or Exponential under Moving Average Calculation.
    11. Click Refresh, then Close.
    12. View the envelope trading band around your price pattern.
    13. Go back to step e) and choose Bollinger Bands on the pull-down menu and repeat the rest of the steps.
20. Summary:
    1. Trends can be plotted on price charts manually to help identify changes in the nature of a price pattern. These changes can serve as powerful trading signals.
    2. Moving averages provide a "better than the eye-ball" approach to identify changes in trends. As trends change, trading signals are generated. Using these trading signals we are more likely to remove emotion. This alone will heighten our likelihood of success.
    3. Moving averages answer some of the oldest truths of successful trading, letting profits run and cut losses short. This forces us to obey the rules by providing specific buy and sell signals based on those principles. However, since they follow trends, they work best when markets trend. They perform less well when markets get choppy and trade sideways.
    4. Trading signals based on moving averages follow the trends, therefore they lag the action. All moving average trading signals will be somewhat late.
    5. The value of the moving average concept is enhanced when trading bands are placed above and below the moving average.
    6. Trading envelopes position lines parallel to the moving average on either side of the MA at some constant percent above & below the MA.
    7. Bollinger Bands are a more sensitive version of the trading band where the width of the envelope depends on the volatility of the stock. The band is self-adjusting being calculated from the standard deviation of the price data.
    8. Moving averages and trading bands find their greatest value with stocks that trend. They are less effective when the price action is sideways. We will cover indicators that work best with side-ways patterns in later Guideposts. It is the combination of them all that will increase our probability of success. Stay tuned.

1. In our discussion last week we found the data spread as important as the average (head in the oven, feet in the freezer, but on average, pretty comfortable!). For instance, 10 prices in a row of $6 each would have an average of $6. Three prices, $3, $6, and $9 would have the same average of $6.

      					              Range
        Average    # Observations  (Data Spread)
     1.    $6           10 	       $0
     2.    $6            3    		   $6
   

2. These two samples are very different, both the data spread and number of observations. The average by itself tells us nothing about the nature of the data other than its location, or where the data is centered.
3. Variability measures data spread. On price charts, the variability is called volatility. This is used in charting services on trading ranges or trading bands. The following discussion will help understand how this takes place.
4. Let's start with the 10 piece sample below to understand how we measure variation:
   
   in the data set (there is one 3, two 4's, etc.), we see how the average represents the "balance" point in the data. That is, there is equal weight on each side of the average.

   Histogram representing the location of the data:
   

5. The easiest way to measure spread is to subtract the smallest value from the largest,

   (8 - 3) = "range" = 5

   While this measures data spread, it is unduly influenced by outliers, or data points at the end. For instance, if point #7 in the data set were 13 rather than 8, the range would be

   Histogram representing the location of the data:
   

6. The range doubled but the average increased less than 10%. Therefore, the range is not a good measure of spread.
7. We can also compare each data point with the average, to measure the spread, subtracting the average from each data set value:
   
8. The sum of differences does not help us measure variability. Rather it gives us zero. By definition, the average is the center. Therefore, there will be as many values above the average (+'s) as below (-'s). That is why the sum of the differences was equal to zero.
9. The sum of the differences was zero because there were as many plus signs as minus signs in the data. To handle this sign issue, we square each value in the Difference column, or multiply each value by itself, to get only positive values (like signs multiplied always give us a positive sign);

           (+ X +)  =  +,       (- X -)  =  +
   

10. So, anything multiplied by itself will always have a positive sign. The "Diff Squared" column provides the result:
    
11. The Diff Squared column now gives a useable measure of the data spread around the average. To finish, we find the average of the Diff Sqd column with a twist: we divide the sum not by 10, but by 9, one less than the total observations. We'll leave "why" to the statisticians.
12. With this new average, we're almost done. The last step is to see that we have taken dollars and multiplied it by itself, so the unit for the 2.49 average is "Dollars squared." We want dollars, not squared dollars. So the last step is to take the square root of the 2.49, giving us 1.57.

                  Standard Deviation (sigma) = sqrt($2.49) = $1.57
    

13. This is what we've been looking for. It is called the Standard Deviation, or Sigma. It is easy to find on a scientific calculator and is widely used to describe the shape of a data distribution.
14. What is important to us is what it is and how to use it. It is not important how it is calculated, even though you understand that now.
15. One final point: The standard deviation is a useful measure of variability for the following reason. If the data were normal, that is symmetrical with most data in the middle and few points on the ends, we can make very definite statements. Such as, 68% of the data will lie within two standard deviations. 95% will lie within two standard deviations. This is why John Bollinger placed his upper and lower trading bands two standard deviations from the center moving average. This means it is not usual for data points to occur near the bands. When a data point does appear near the band, it signifies an unusual, important departure from the norm.
    

May 7 - Oscillators [Leading indicators]

May 14 - Wilder's relative strength index (RSI) Divergences, Convergences

May 21 - Stochastics, MACD, On Balance Volume (OBV)

May 28 - TC-2000 proprietary indicators (BOP, MS, TSV)

We provide TC2000 tutorials to members upon request.