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]
63 Chapter 21: Normalizing Indicators volume. Other than that, no general guidance is possible due to variation in indicator parameters and formulations. However, an example of a setup that works quite well is a 14-day RSI with 50-day, 2.1-standard deviation Bollinger Bands. Using this combination of parameters, overbought and oversold levels for most stocks are easily identified, and clear divergences are delineated at many turning points. A sharp eye will soon hone in on the appropriate Bollinger Band parameters for any indicator application. Start with the same average-length-picking approach that was discussed earlier for stocks. What you want to see is a slow transition in which the average lives in the upper range of the indicator during uptrends and in the lower range during downtrends. However, the average should stay in the middle portion of the chart, for MFI roughly between 25 and 75, and for RSI a bit tighter, between 30 and 70. If the average gets into the upper quarter or lower quarter of the range, you have an average that is too short. If the average barely budges from the midpoint, you have an average that is too long. Then set the number of standard deviations for the bands, starting with 2, so that between 85 and 90 percent of all observations fall within the bands. There is a reason that these indicator parameters vary greatly; indicators tend to be distributed in quite a different manner than stocks. In fact, some indicators are distributed in a decidedly non-normal fashion. Stochastics tends to have fat tails and can even have a U-shaped distribution where the tails are fatter than the middle (Figure 21.1), while RSI tends to have thinner tails. However, you need not concern yourself with the statistics. If you follow the above procedure, youll end up with a workable approach. Table 21.1 Trial Bollinger Band Values for Indicators Indicator | Length | Width | 9-period RSI | | | 14-period RSI | | | 10-period MFI | | | 21-period II | | |
Part V: Advanced Topics  0.5 D.f. 07 Figure 21.1 Distribution chart, 10-day stochastic, IBM, two years. This is about as far from a normal distribution as you can get. Now, treat the upper band as you would an overbought level; for example, 70 for the Relative Strength Index. And treat the lower band as an oversold level; for example, 20 for the Money Flow Index. If you have set the correct parameters, all the regular decision rules will hold, such as buying a positive crossing of the lower band or treating a tag of the upper band as an overbought signal. There is no need to worry about rigid frames or rules any more; they have been made moot. Trade in sync with the prevailing trends using the indicators within the bands as your guide to opportunities. And now for our neat trick. If you will, a bit of indicator magic: %b is normally used to depict the location of a data point, typically the close or last, within the Bollinger Bands. At 1.0 we are at the upper band, at 0.5 we are at the middle band, and at 0.0 we are at the lower band. The range of %b is not confined to the zero-to-1 interval. A reading of 1.1 tells us we are 10 percent of the BandWidth above the upper band, and a reading of -0.15 says we are 15 percent below the lower band using BandWidth as a gauge. First, calculate and plot your indicator. Second, plot Bollinger Bands on it using the method just described to set the parameters.
Chapter 21: Normalizing Indicators Table 21.2 Normalized Indicator Formula (Indicator - indicator lower band)/(indicator upper band - indicator lower band) Third, calculate %b (see Table 21.2) using the indicator and the bands just plotted. Fourth, plot %b alone as a normalized version of the indicator! Ta-da! See Figures 21.2 and 21.3 for examples using MFI. What we have done is redrawn the indicator using the upper and lower bands as the boundaries instead of the absolute possible range of the indicator-0 to 100 in the case of RSI. This normalization of indicators is one of the most important uses of the %b formula. To refer to an indicator normalized in this fashion, we write %b(RSI) (see Figures 21.4 and 21.5). Normalization of an indicator using Bollinger Bands allows the indicator to be included in trading systems in an adaptive manner. The technique can be applied to open-ended indicators that  Figure 21.2 MFI with Bollinger Bands, Dupont, 150 days. The bands define high and low rather than the rigid 20 and 80 levels typically used.
[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]
|
Menu
