Time Cycles
Previous  Top  Next

You can use Time Cycle studies to identify patterns in the time element of price bars. In the more common types of chart (such as Candlestick and OHLC), price bars display information about both time and price. Each bar displays the important prices that occurred during a specific time period (such as a day, a week, a month, an hour, five minutes, fifteen minutes, etc). When significant changes in price occur at regular intervals, Time Cycles provide a way to study that phenomenon, especially when a market has been seen to follow more than one cycle and when multiple cycles indicate significant events at about the same time. See Disclaimer.

Fibonacci vs Fixed

ChartOverlay lets you explore two varieties of Time Cycles. Fixed Time Cycles have fixed intervals (eg, every 21 bars, or every 89 bars); because all intervals in a Fixed Time Cycle item have the same size, the ratio of any interval to its predecessor is 1.0. Fibonacci Time Cycles have increasing intervals, where successive intervals are taken from the Fibonacci Sequence (1, 1, 2, 3, 5, 8, 13, ...). As the following table shows, from 8 onwards, each new Fibonacci Time Cycle interval is approximately 1.6 times greater than its predecessor (eventually settling at 1.618). It is this ratio, not the individual interval sizes, that is of interest for Fibonacci Time Cycles.

Line #
1
2
3
4
5
6
7
8
9
10
11
12
...
Interval
0
1
1
2
3
5
8
13
21
34
55
89
...
Distance
0
1
2
4
7
12
20
33
54
88
143
232
...
Ratio


1.000
2.000
1.500
1.667
1.600
1.625
1.615
1.619
1.618
1.618
1.618

Line # - a simple sequence number
Interval - the number of bars between this line and its predecessor (the Fibonacci sequence)
Distance - the number of bars between this line and the first line (the sum of the intervals)
Ratio - this interval divided by its predecessor (the Fibonacci ratio)

Drawing a Time Cycle Item (Standard Method)

When you use the Standard Method, you tell ChartOverlay to draw Time Cycle lines at certain intervals (either Fixed or Fibonacci). These intervals will be measured in Price Bars (remember that each Price Bar actually represents both Time and Price).

Begin by pressing the btntime button or by choosing Time Cycles from the menu.

Then choose either Fibonacci or Fixed in the dialog box that appears. If you choose Fixed, you must also specify the length of your interval (ie, the number of price bars in each cycle). If you choose Fibonacci, the interval sizes will be determined automatically by the Fibonacci Sequence.

Specify a Calibration Length (see below).

Select any other options you would like , such as Line Color and Line Style.

·Draw a Calibration Line, in response to which ChartOverlay will draw the Cycle Lines.  

Calibration Lines

Because ChartOverlay cannot actually see the data in your underlying chart, it cannot count bars unless you give it some help. You can do this by drawing a Calibration Line that spans several bars on your chart and by entering the number of bars that were spanned (the Calibration Length) in the Time Cycle dialog.

ChartOverlay uses these two items, the Calibration Length, which is given in bars, and the Calibration Line, which can be measured in screen pixels, to determine the width of a single bar (including the space separating it from its neighbor). This can be calculated by dividing the length of the Calibration Line by the Calibration Length. Given the average width of a single bar (plus spacing), ChartOverlay can then determine the width of N bars. This is analogous to measuring a picket fence by counting pickets.

It is important that the Calibration Line begin and end on corresponding points on the first and last bar (eg, both leading edges, or both trailing edges). This ensures that the Calibration Line spans exactly N bars and N spaces.

In the following examples Calibration Lines A, B and C are valid; line D is not. Note that each of lines A, B and C begins and ends at the same relative position of its beginning and ending bars (left edge, center and right edge, respectively). If the beginning and ending bars have the same width, as appears to be the case here, then Lines A, B and C will each have the same length, which will correspond to the width of 12 bars and the width of 12 spaces. Dividing the length of any of these three lines will yield the average width of one bar and one space. Line D is invalid because it starts at the left edge of one bar and ends at the right edge of another; it is not the same length as the other lines because it spans 13 bars and 12 spaces.

CalibrationLines

To understand why the Calibration Length should be greater than one, consider the following example. Suppose you are looking at a chart that is displaying 100 bars and suppose the width of your chart is 450 pixels. Then each bar (including the spacing between bars) will occupy an average width of 4.5 pixels, which means that most bars will be five pixels wide while about one in ten will be only four pixels wide. Thus, the longer your Calibration Line, the more accurate your result (but the more difficult it will be to draw the line).

An Alternate Method

The Alternate Method is very similar to the Standard Method, except that it does not employ the precision counting of bars. Instead, you let the data dictate what you draw and you do not worry about whether the intervals correspond precisely to bar widths.

As with the Standard Method, you draw a Calibration Line, but in this case Calibration Length is arbitrary (bigger numbers are easier to work with; experiment). Once you have created your Calibration Line, you then adjust its end point (ctrl-click the end point and drag it) while watching where the Time Cycle lines are drawn as you make the adjustment. Note that you can also ctrl-click and drag the Calibration Line to move it and the Time Cycle lines together.

You can use this method to feel out the data while looking for cycles. If it happens that the Calibration Line touches the two bars at similar points, you can count the bars spanned by the Calibration Line and enter that value in the dialog to get accurate numbering. It is more likely, however, that the two bars will be touched at dissimilar points, in which case the numeric labels attached to the cycle lines will be meaningless.

Display

There are several points worth noting in the following example.

TimeCyclesEx
There are three examples of Time Cycle items in this chart:
·Blue - Fibonacci Cycles drawn using the Standard Method with a Calibration Length of 12. Note that the Calibration Line ends at the cycle line labeled 5. This is because 1 + 1 + 2 + 3 + 5 = 12; the label numbers indicate the intervals (the number of price bars) between lines.  
·Red - Fibonacci Cycles drawn using the Alternate Method. The length of the Calibration Line is not important.  
·Cyan - Fixed Cycles with a Calibration Length of 12 and a Cycle Length of 21. There is no relationship between the Calibration Length and the Cycle Length. The Calibration Length tells ChartOverlay the length (in bars) of the Calibration Line and the Cycle Length tells it how many bars to count before drawing a cycle line.  

The labels for Fibonacci Cycle Lines tell you the distance (in bars) between the line with the label and the line to its left (some lines may be invisible, as described below, but they still affect where subsequent lines are drawn). The labels for Fixed Cycle Lines tell you the distance (in bars) from the very first line.

Note that Time Cycle labels appear at the top of the window. This is to prevent them from conflicting with the labels for Fibonacci Time Extensions. When one of these numeric labels would overlap another (for example, for the first few Fibonacci Cycle Lines), it is offset slightly below its normal position; if a number would require a second offset line, it is simply not displayed. When there is more than one Time Cycle item on a chart, each has its own row of labels (or pair of rows), as seen here.

When the space between two cycle lines for the same Time Cycle item is too small (less than 10 pixels), the newer line will be invisible, even though its label may be visible. This avoids crowding for the smaller intervals of a Fibonacci Time Cycle item. Even though a line is not visible, subsequent lines will be drawn as if the missing line were visible.

Note how often the lines of the two Fibonacci Time Cycle items overlap. This occurs because the ratio between consecutive numbers in the Fibonacci sequence tends to be constant (approximately 1.618..., see table above).

Note that when using the Alternate Method, when the Calibration Line is short (as it is here), it is difficult to exert precise control over the cycle lines to the right. This is because moving the end of the Calibration Line by a single pixel may move lines further to the right be several pixels. Using a longer line with a bigger length makes drawing easier; experiment.

Example

This example, created by Don McCullar from Denton, Texas, while using StockCharts.com®, is a simplified version of an example presented in a previous section. Note how the Fibonacci Time Cycles often coincide with reversals in the price chart. The Calibration Line runs from 4/26 to 5/25, which represents 21 trading days (not counting 4/26 itself). Using this line in conjunction with a Calibration Length of 21 yields a series of lines at the distances indicated by the Fibonacci sequence.

dm spx 20100722a2