back start next


[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [ 24 ] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81]


24

Part II: The Basics

Table 7.2 Recommended Width Parameters for Bollinger Bands

Periods

Multiplier

platform, for example. But none is more important than the evolution of the markets. The initial Bollinger Band parameters were developed almost 20 years ago, and the markets have changed dramatically since then. For example, stock index futures were new, unproven vehicles at the time. There can be no doubt that the markets have evolved since then, and our approaches need to evolve as well.

To summarize the findings: At 20 periods and 2 standard deviations you get containment between 88 and 89 percent in most markets. To keep that containment percentage constant when you shorten the calculation period to 10 days, you need to decrease the bandwidth from 2.0 to 1.9; and when you lengthen the calculation period to 50 days, you need to increase the bandwidth from 2.0 to 2.1.

For calculation periods less than 10 or greater than 50, changing the periodicity of the bars is more appropriate. For example, if you require a shorter calculation period than 10 days, a shift to hourly bars might be better than trying to squeeze the calculation period ever tighter. There are seven trading hours in an NYSE day; the first half hour from 9:30 a.m. to 10:00 a.m. should be counted as an hour. So 35 hours is equivalent to 5 days. As a general rule, try to keep the calculation period near 20 or 30, the ranges within which there is a lot of experience. Thaf s better than trying to push the envelope and getting unexpected results.

Why a simple moving average? For years a father and son team advertised "better Bollinger Bands" in Investors Business Daily. Their "secret"? They used an exponential moving average as the measure of central tendency. Yet this book still recommends a simple moving average. The reason is that a simple moving average is what is used in calculating the volatility used to set the bandwidth, so it is internally consistent to use the same



Chapter 7: Construction

average to set the center point. Can you use an exponential average? Of course. Any average will work. But in doing so you are introducing an extraneous factor that you might or might not have to pay attention to. In our testing, no clear advantage was conferred by using an exponential or front-weighted average, as you can see if you compare Figures 7.1 through 7.3. So in the absence of a compelling argument, you should stick to the simplest and most logical approach.

What about mismatching the calculation periods? One popular mismatch is using longer periods for volatility and shorter periods for the average. The idea is to capture information from the dominant volatility cycle for the bandwidth while using the best measure of trend for the midpoint-for example, using a 50-period moving average for the center point and a 20-period volatility for the bandwidth (Figure 7.4). This is done less than swapping the average types in our experience, but it is done. I frankly cant see why you would want to introduce yet another variable to an already complex process, but if it floats your boat...

One last variation that is quite popular is to deploy multiple bands at the same time. This can be done in two ways. One is to

7/00 8/00 9/00 10/00 11/00 12/00 1/01

Figure 7.1 Bollinger Bands, 20-day simple moving average, Deere & Co., 150 days. Classic Bollinger Bands.



Figure 7.3 Bollinger Bands, 20-day front-weighted moving average, Deere & Co., 150 days. Front-weighted is even faster.



[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [ 24 ] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81]