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] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] [183] [184] [185] [186] [187] [188] [189] [190] [191] [192] [193] [194] [195] [196] [197] [198] [199] [200] [201] [202] [203] [204] [205]


27

The single moving average at day * is Smoothing the new series Af gives

and the third smoothing gives

Jack K. Hutson, -Good TRIX." TetJmiaU Analysis of Stocks & Con

M - -

(f-t + Bf-s+fifi. ., + 7f3 + 6fi. ; + 3/-,-ff) 27

The effect of the weighting puts increased importance on the middle values, reducing the significance of both ends. With exponential smoothing as well, the increased emphasis will be placed on the n - 2 value, the price which is repeated most often.

The Parabohc Time/Price Sjstem

One of the primarj complaints about trend-following sjstems is that the built-in lag destrojs the trade. In the Parabolic Time/Price System, Wilder has reduced the tag by increasing the speed of the trend (shortening the dajs) whenever prices achieve new profitable levels. The philosopliy of the Parabolic Sjstem is that time is an enemy. Once a position is entered, it must continue to be profitable or it will be liquidated.

The Parabolic Time/Price System is ahvajs in the marlcet; whenever a position is closed out, it is also reversed The point at which this occurs is caned the Stop and Reverse (SAR). When plotted, it is similar to a trendline, although it has a decreasing lag (the distance between the current price and the trendline, or SAR point), as shown in Figure 5-4.

To calculate the SAR value, fast assume a long or short position. If the market has recently moved lower and is now above the lows of that move, assume a long. Call the lowest point of the previous trade the SAR initial point (SIP), because it will be the starting point for the SAR calculation (SAR = SIP).

Calculate each following SAR as

SAR(tomorrow) = SAR( prior) + AF x [High(today) - SAR( prior)

This is an exponential smoothing using the prior high price. It is unique because the smoothing constant AF, called the acceleration factor, changes. AF is set to .02 at the beginning of the trade. Each day that a new extreme occurs (a new high when long, or a new low

3 J Welles Wfil.ler, Jr , New r.i.ceftr in TecLuical Tradiug cy.-ten, (TreudEesearcL, ieensboro, IT?, 1S7B)

FIGURE 5-4 Parabolic Time/Price Sjstem.



\ SARshort \> (uses lows)

when short), AF is increased by .02. In moving average dajs, AF begins at 99 dajs and increases in speed to a maximum of a 9-day moving average, but not in a linear faion.

In the SAR calculation, High(today) is used when a long position is held, and Low(today) is used when a shorl is held. An additional rule also prevents premature reversal by not allowing the SAR to enter the price range of the mostrecent2dajs.

If long, the SAR may never be greater than the low of today or the prior day. if it is greater than this low, set the SAR to that low value. A reversal will occur on a new intraday low that penetrates the SAR.

If short, the SAR is calculated in the same way based on the highs.

Winers Parabolic Time/Price System was a new and interesting idea-varjing the exponential smoothing constant. This could also be accomplished by varjing the moving average dajs, but not as effectively. The constant increment of .02 and the limitations of the range, .02 to .20, are likely to be shortcomings. Traders might want to relate the changing speed to more basic price phenomena such as volatility or momentum. For those readers interested in the fiirther development of this method, the Parabolic Time/Price Sjstem is combined with Directional Movement (described in Chapter 23) to form the Directional Parabolic Sjstem discussed in Chapter 6.

The MasterTrading Formula

Another method of variable speed is Marts The Master Trading Formula. it uses an exponential smoothing formula in which the smoothing constant and band are calculated daily, based on market volatility. Mart averages the extended daily range (using the maximum of yesterdays close and todays high and low) and the net price movement over the same interval to define a correlated volatility factor. This value is then used to look up a smoottiing constant in a modified log-scale table, in which an 8°o smoothing (24-day equivalent) is used for the lowest volatility and a smoothing (2-day equivalent) for the greatest volatility, with the median value at (13-day equivalent).

A band is also placed around the trendline, based on an inverse relationship to the correlated volatility factor The band is widest when there is low volatility and the smoothing constant produces a slow trend; it is narrowest in fast, highly volatile markets.

Rules for trading Marts Master Formula are centered around the basic band trading rules. The sjstem is alwajs in the market, taking long positions when the upper band is broken and short positions when the lower band is crossed.

Raschkes First Cross

Although we usually trade a trend from the beginning to end based on some smoothing method, Linda Bradford Raschke has shown that a selected piece of the trend move can be a very reliable trade. Her First Cross sjstem enters a trend trade on the first pullback after an initial trend signal based on a faster momentum indicator. She did not offer any exit rules, but there are some that can be suggested. To create this program we need



slowMA Slow-moving average, part of oscillator calculation festMA Fast-moving average, part of oscillator calculation osc Moving average oscillator, fastMA - slowMA

trend Moving average of osc

The oscillator is the difference between the fest- and slow-moving averages osc = festMA - sIowMA

Donald S M.nt. The Tra.lmg F. tniula Wmsor Bo. .k.-, Bnghtwaters, 1 , 19B1)

To generate a buy or sell entry signal, use the following steps (comments on the right):

Bl osc[l]> trerd and osc trena osc crosses treio movingdown B2, loh > low[l] lowofbarnses

B3 buy

51. osc[l] < trend aro osc >= trero osc crosses trend moving up

52. nigh < highfl] high cf the bar declines

53. sell

A likely exit rule for the First Cross sjstem should maintain the short-term nature of the selection. You might simply consider the entrj on the first pullbad: as better entry timing for a longer-term trend. Or, you might want to trade only the short period isolated by the first pullback until this initial impulse is exhausted. If you chose the second method, you might try the following exit condition:

Exit for a long position

XI. closeLl] > clo5e[2] close of bar turns up (confirms trend) X2. close < closeLU after going up the close turns down

X3. RXTt long

Raschkes idea is an excellent exanple of selectivitj- First, you recognize that the beginning of a trend is different from other times. As traders recognize a change of direction there is an increasingly strong move. Because most sjstems lag the maiket, they are often too late or miss the first part of the turn, which could create large profits. As an alternative, this technique waits for the first move to be exhausted, then enters in anticipation of another fasl surge as the new trend reasserts itself, once these early moves are over, the general trend move may not be as easj to work with, and you may find Yourself trjing to enter as the trend comes to an end.

COMPARISON OF MAJOR TREND SYSTEMS

A section on u-ends cannot be complete without looking at the comparative performance of the six most popular approaches:

1. Simple moving average

2. Rxponential smoothing

3. Linear regression slope

4. N-day breakout

5. Swing breakout

6. Point-and-figure

Trading Rules

To see the characteristics of each sjstem, the trading rules will be as simple as possible. All six sjstems are ahvajs in the market. That is, once they enter a long or short position, they ahvajs reverse on a new signal, going from long to short, or short to long. There are no stop-losses or other rid; control rules; therefore, each sjstem has its own.



[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] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] [183] [184] [185] [186] [187] [188] [189] [190] [191] [192] [193] [194] [195] [196] [197] [198] [199] [200] [201] [202] [203] [204] [205]