Our second Bollinger Band® demonstration method relies upon the idea that strong price action accompanied by strong indicator action is a good thing. It is a confirmation approach that waits for these two conditions to be met before giving an entry signal. Of course, the opposite, weakness confirmed by weak indicators, generates a sell signal.
In essence this is a variation on Method I, with an indicator, MFI, being used for confirmation and no requirement for a Squeeze. This method may anticipate some Method I signals.
We'll use the same exit techniques, a modified version of Parabolic or a tag of the Bollinger Band on the opposite side of the trade. The idea is that both %b for price and MFI must rise above our threshold. The basic rule is: If %b is greater than 0.8 and MFI(10) is greater than 80, then buy.
Recall that %b shows us where we are within the bands; at 1 we are at the upper band and at 0 we are at the lower band. So, at 0.8 %b is telling us that we are 80% of the way up from the lower band to the upper band. Another way of looking at that is that we are in the top 20% of the area between the bands. MFI is a bounded indicator running between 0 and 100. 80 is a very strong reading representing the upper trigger level, similar in significance to 70 for RSI.
So, Method II combines price strength with indicator strength to forecast higher prices, or price weakness with indicator weakness to forecast lower prices.
We'll use the basic Bollinger Band settings of 20 periods and +/- two standard deviations. To set the MFI parameters we'll employ an old rule; indicator length should be approximately half the length of the calculation period for the bands. Though the exact origin of this rule is unknown to me, it is likely an adaptation of a rule from cycle analysis that suggests using moving averages a quarter the length the dominant cycle. Experimentation showed that periods a quarter of the calculation period for the bands were generally too short, but that a half-length period for the indicators worked quite well. As with all things these are but starting values. This approach offers many variations you can explore. Also, any of the inputs could be varied as a function of the characteristics of the vehicle being traded to create a more adaptive system.
Table 19.1 - Method II Variations
Volume-Weighted MACD could be substituted for MFI.
The strength (threshold) required for both %b and the indicator can be varied.
The speed of the parabolic also can be varied.
The length parameter for the Bollinger Bands could be
The main trap to avoid is late entry, since much of the potential may have been used up. A problem with Method II is that the risk/reward characteristics are harder to quantify, as the move may have been underway for a bit before the signal is issued. One approach to avoiding this trap is to wait for a pullback after the signal and then buy the first up day. This will miss some setups, but those remaining will have better risk /reward ratios
It would be best to test this approach on the types of stocks you actually trade or want to trade, and set the parameters according to the characteristics of those stocks and your own risk/reward criteria. For example, if you traded very volatile growth stocks you might look at higher levels for the %b (greater than one is a possibility), MFI and parabolic parameters. Higher levels of all three would pick stronger stocks and accelerate the stops more
quickly. More risk adverse investors should focus on high parabolic parameters, while more patient investors anxious to give these trades more time to work out should focus on smaller parabolic constants which result in the stop-out level rising more slowly.
A very interesting adjustment is to start the parabolic not under the entry day as is common, but under the most recent significant low or turning point. For example, in buying a bottom the parabolic could be started under the low rather than on the entry day. This has the distinct advantage of capturing the character of the most recent trading. Using the opposite band as an exit allows these trades to develop the most, but may leave the stop uncomfortably far away for some.
This is worth reiterating: another variation of this approach is to use these signals as alerts and buy the first pullback after the alert is given. This approach will reduce the number of trades--some trades will be missed, but it will also reduce the number of whipsaws. In essence this is quite a robust method that should be adaptable to a wide variety of trading styles and temperaments.
There is one other idea here that can be important: Rational Analysis. This Method buys confirmed strength and sells confirmed weakness. So wouldn't it be a good idea to presort our universe of candidates by fundamental criteria, creating buy lists and sell lists? Then take only buy signals for the stocks on the buy list and sell signals for the stocks on the sell list. Such filtering is beyond the scope of this book, but Rational Analysis, the juncture of the sets of fundamental and technical analysis, offers a robust approach to the problems most investors face. Prescreening for desirable fundamental candidates or problematic stocks is sure to improve your results.
Another approach to filtering signals is to look at the EquityTrader.com Performance Ratings and take buys on stocks rated 1 or 2 and sell on stocks rated 4 or 5. These are front-weighted, risk-adjusted performance ratings, which can be thought of as relative strength compensated for downside volatility.
- The method buys strength
- Buy when %b is greater than 0.8 and MFI is greater than 80
- Use a parabolic stop
- May anticipate Method I
- Explore the variations
- Use Rational Analysis