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]


135

iMuinpie time penoas : Linear Kegression b\cf>e + biocnastict

(period = length of exponential trend datal = shorter time period of stochastic data2 = longer time period of linear regression slopel

input: fast(5), 5low(50);

vars: slope(O), overbot(80), oversold(20), stoctitO);

(calculate stochastic over shorter data per-odl stoch @slowk(fast) oi datal;

(Linear regression slope)

slope = LinearRegSlope(close of data2.slow);

if slope 0 then exitshort on close; if s)ope < 0 then exitlong on close-,

(long signal : slope must be up and stoctiastic below low ttiresiihold 1 if slope > 0 and stoch < oversold then buy on close;

Ishort signal : slope must be down ard stochastic above upper threshholc if slope < 0 and stoch > overbot then sell on close;

ELDERS TRIPLE-SCREEN TRADING SYSTEM

The Triple-Screen method combines three time frames to remove the disadvantages of eadi one. It combines indicators that are both trend-following with oscillators (normally associated with a countertrend direction). Dr. Elder has observed that each time frame relates to the next by a factor of 5, That is, if you are using daily data as the middle time period, then the shorter interval will be divided into five parts, and the longer period win be 5 days or I week.

To be practical, it is not necessary to divide a 6-hour trading day into five intervals of I hour and 12 minutes Rounding to I hour is close enough. If, for example, you want to focus on frading a lOminute chart, then the middle interval is 10 minutes, the short-term is 2 minutes, and the longterm is I hour (not 50 minutes).

In the following description, screen I holds the longest time frame while screen 3 shows the shortest one.

Screen I: The Major Movement

The long-term view gives the market tide, a clear perspective of the major market frend, or sometimes lack of frend. Weekly data is used for this example, which is consistent with most experience that less frequent data (Le., weekly or monthly) smooths the price movement by eliminating interim noise. Although there are many other choices for a long-term frend, the Triple-Screen approach uses the slope of the weekly MACD, where the histogram thai represents the MACD value is very smooth, equivalent to, for example, a 13-week exponential smoothing. The frend is up when the MACD bar, or 13-week exponential value is higher than that of the previous week; the frend is down when this weeks value is lower.

TfmDr .1 1.1 Tra.lmgfor aLiviue (J-hu Wiley & Sons 1

Screen 2: The Intermediate Movement

Using an oscillator, the Screen 2 identifies the time period in which we would trade. Again, the specific oscillatcr is not as important as the time frame and the ability to identifj market waves in the major movement of Screen I. Two oscillators are suggested, the Force Index and Elder-ray, both described below A stochastic can also be used.

1. Force Index

Force Index = volume(today) x [close(today) - close (yesterday)

The Force Index is then smoothed using a 2-day exponential smoothing, which has a smoothing constant of -333,



and the resulting value is used to determine the overbought and oversold levels.

Entering a long position using the Force Index is not as clear as when using the Elder-ray; however, the following steps are necessarj-

Step 1. The trend in Screen 1 is up.

Step 2. The 2-day exponential of the Force Index falls below its centeriine and does not fall below the multiweek low. When using a stochastic instead of the Force Index, buy when the stochastic falls below 30.

2. Elder-Ray

The Elder-ray is a technique for separating bullish and bearish movement.

Bull power high - 13-day exponential smoothing

Bear power low - 13-day exponential smoothing To determine when to buy using the Elder-ray and Screen 1, the following twost s are necessarj:

Step 1. The trend in Screen 1 is up.

Step 2. Bear power is negative but rising., bear power must not be positive.

Two additional steps may be used to filter trades and improve performance, but are not required:

Step 3. The last peak in bull power is higher than the previous peak (the most recent bull power should not be significantly lower than the previous peak).

Step 4. Bear power is rising from a bullish divergence.

The opposite rules apply for sell signals.

Screen 3: Timmg

The final screen is for fastest response, primarily for identifjing intraday breakouts. To improve the point of entry. Screen 3 can be used to set long positions when the current price moves above the previous days high. There is no calculation involved, simply a Buy Stop order using the shortest time period. For this example, where Screen 1 is weekly and Screen 2 is dally. Screen 3 would be hourly. Then to get a new buy signal, the hourly bar must move above the highest hourly bar of the previous day.

Stop-Loss

Every sjstem needs rid; control, and that most often comes as a Stop-Loss order. The Triple-Screen approadi positions the stop as a three-step process. For a long position.

Step 1. First place the stop below the low of the day of entry, or the previous days low, whichever is lower.

Step 2. Move the stop to the break-even level as soon as possible. Naturally, there must be some room between the stop-loss level and the current price; otherwise, the stop will alwajs be hit

Step 3. Move the stop to protect 50°o of the highest profits. In addition, you may consider taking profits when the stochastic or Force index moves above the 70°o level.

ROBERT KRAUSZS MULTIPLETIME FRAMES

Although multiple time frame techniques have become more visible, the most robust approadi has been taken



by Robert Krausz. To understand the importance of first arriving at a sound theory before implementing and testing a trading program, we need to briefly review the characteristics of performance that indicate a robust method.

When testing a trend-following sjstem,we should expect that a trend of 100 dajs, compared with atrend of 50 dajs, will produce larger profits per trade, greater reliability, and proportionally fewer trades. As you increase the calculation period, this pattem continues; when you reduce the calculation period this pattem reverses. You are prevented from using very short calculation intervals because slippage and commissions become too large; the longesl periods are undesirable because of large equity swings. There must be a clear, profitable pattem when plotting retums per trade versus the average holding period.

The sophistication in Krauszs work lies in his understanding of this pattern, and its incorporation into the stmcture of his program. The Fibonacci Trader. Krausz works in three time franes rather than two. Each time has a logical piwpose and is said to be modeled after Ganns concept that the maitets are essentially geometric. The shortest time frame is the one in which you will frade; in addition, there are two longer time frames to put each one into proper perspective. The pattems common to time fines are easily compared with fractals; within each time frame is another time with very similar pattems, reacting in much the same way. You cannot have an hourly chart without a 15-minute chart, because the longer time period is composed of shorter periods; and, if the geomdr holds, then characteristics that work in one time , such as siport and resistance, should work in shorter and longer time frames. Within each time there are unique levels of support and resistance; when they converge, the chance of success is increased. In Krauszs work, the relationships between price levels and profit targets are woven with Fibonacci ratios and the principles of Gann.

One primary advantage of using multiple time frames is that you can see a pattem develop sooner. A frend thai appears on a weekly chart could have been seen first on the daily chart. The same logic follows for other chart formations. Similarly, the application of pattems, such as support and resistance, is the same within each time frame. When a support line appears at about the same price level in hourly, daily, and weekly charts, it gains importance.

As a well-known trader,3 Krausz brings more than just three time frames and some unique sfrategies to one display screen. He endows the program with six rules:

Laws of Multiple rime Frames

1. Every time frame has its own stmcture.

2. The higher time frames overrule the lower time frames.

3. Prices in the lower time frame stmcture tend to respect the energj points of the higher time frame shncture.

Robert EraiiEZ. WD . Tre.nire Discovered (Ge4iietnc Trailers bstitute. TS"? £E Pth street, cmte 272. Ft Laii.lefUle. FL 33311 Also see references in the

ction "Fibonacci andHuman Behavior." h.ler 14 1 See jack Schwr. The New bbukt Wiz.-u.L- (J..hii Wiley & Sons. New York, IF, 19 . i - -HEobertrj-.msz

4. The energj points of siq3port/resistance created by the higher time frames vibration (prices can be validated by the action of the lower time periods.

5. The trend created by the next time period enables us to define the tradable trend.

6. What appears to be chaos in one time period can be order in another time period.

Using three time frames of about the same ratio to one another (lO-minute, 50-minute, and daily), with daily being the longest. Figure 19-2 shows the June 98 contract of U.S. bonds with a number of techniques applied over multiple time frames. Figure 19-2a uses only daily bars, while Figure I9-2b has lO-minute bars, both charts drawn on the same price scale to facilitate comparison.

FIGURE 19-2 Krauszs multiple time frames, Jun 98 US. bonds, (a) Daily chart with Gann swings and a stepped moving average, (b) Multiple time frame stmcture for the corresponding 4 dajs.



[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]