Articles
TD Sequential And Ermanometry For Intraday Traders By Andrew Coles
You can automate time-based techniques to apply to intraday charts. In the first part of this series, we discussed Fibonacci’s bar-counting technique and the Lucas number series. This time, we will examine two others: the TD Sequential and Ermanometry.
Read First Article: Automated Techniques For Intraday Traders
Time-based techniques are effective and can be applied to intraday trading, but unless they are automated, they require a great deal of tending and tracking on the part of the trader. The Fibonacci and Lucas techniques, which we explored in the previous issue, can be mechanized (as seen in that article), but they will not continue to plot without further input. This is also true of William Erman’s Ermanometry, which we will look at later in this article. Meanwhile, Thomas DeMark’s TD Sequential can be automated without requiring further input, as we will see here.
TD SEQUENTIAL
In his book Day Trading Options as well as in various articles, Thomas DeMark demonstrated that his TD Sequential and TD Combo bar counting techniques are equally effective on intraday charts. However, there are many trading platforms that have not coded the indicators, and their manual calculation still presents a challenge for a busy daytrader.
It’s beyond the scope of this article to discuss DeMark’s two indicators in detail. Suffice it to say that TD Combo anticipates price exhaustion within trends, as does the countdown component of the TD Sequential. The setup component of TD Sequential focuses on the momentum component to define price ranges. Thus, if a directional move continues after the momentum component is complete, TD Countdown and TD Combo can be applied to identify areas of trend exhaustion.
However, as Jason Perl stresses in his recent book about DeMark’s indicators, the momentum component is still important, since how the market responds to a setup determines whether it is range-bound. Each time price completes a setup, the price extremes are identified by a support and resistance line known as the TD Setup Trend (TDST). If price fails to break through the TDST, there is an opportunity to take a contrary position insofar as the market doesn’t have the momentum to start the countdown phase — that is, to break through the TDST into a new distinct trend. This means that a mechanized version of the setup is capable of creating many interesting signals of its own.
A TD Sequential buy setup adheres to the following two conditions:
- A bearish price flip whereby price must close higher than the close four bars earlier, followed by a close less than the close four bars earlier.
- Nine consecutive closes, each one less than the corresponding close four bars earlier (where the bar on which the bearish price flip occurs qualifies as bar of the buy setup).
For a sell setup, these conditions are reversed. These two conditions can be programmed into MetaStock and the result is a mechanized version of the setup component of the TD Sequential. See, “MetaStock Code For Intraday Application Of TD Setup.”
Figure 1 is a 30-minute chart of the British pound December 2010 futures. In the main pane, there is a red Fibonacci sequence running from the low at the red arrow, and the blue spikes further along are the TD Sequential Setup. Since the countdown and combo aspects aren’t coded, other potential signals are overlooked; nonetheless, all but the penultimate blue spike capture a market turning point. The middle pane shows another Fibonacci sequence launched from the dark red arrow and the top pane shows a third Fibonacci sequence launched from the green arrow. As the chart is being printed, the dark red sequence has just fired off on the 55th bar as the sharp downtrend halts on a hammer candlestick; at the same time, the TD setup is in the process of firing off. The black rectangles indicate areas of overlap between the TD Countdown and the Fibonacci series. The combined red series plus the TD setup produce some impressive signals, including the two largest moves of the day in the case of the TD setup.
FIGURE 1: 30-MINUTE CHART OF BRITISH POUND DECEMBER 2010 FUTURES. Here you see a red Fibonacci series and DeMark’s setup in blue. The upper panes contain further Fibonacci series launched from more recent turning points.
There is no Lucas series in this chart because it is intended to show how multiple Fibonacci series can be plotted from multiple turning points as the market moves forward. DeMark’s indicators are easier to plot because they do not require a starting point to be input by the trader. Because of this, DeMark’s indicators would be left running in the background, whereas it is up to the trader’s discretion from which points Fibonacci and Lucas series should be launched. As I mentioned in part 1, you can drop down to lower-level charts such as the five-minute to work with the consequences of these signals. But the need to find higher time frame charts is less important when using the TD setup.
If you are wondering about the major swing high highlighted by the magenta circle, note that it was captured by both a Fibonacci series and a Lucas series. However, neither was plotted on this chart and its identification was retrospective. Thus, this chart highlights once again that while most pivots do seem to be genuinely captured by Fibonacci and/or Lucas series, it is impossible to launch a series from every pivot in real time, and consequently, some major price turning points will only be identified retrospectively (however, this doesn’t apply to the DeMark indicators). This highlights again the critical need for price-based technical analysis.
ERMANOMETRY
Our fourth and final intraday mechanized system is based on William Erman’s Ermanometry. Erman’s work on market timing has been published in this magazine, and here I follow his first article from 1999 to create this mechanized intraday timing system. See “MetaStock Code For Intraday Application of Ermanometry.”
In that article, Erman stated that the largest time unit for his approach is the trading day, while stressing that a unit as small as 15 minutes could also be used. With easier accessibility now to intraday charts and with a mechanized version of this timing system, I have found that time units as low as one-minute bars can be used, which is to be expected, given the success of other fractal approaches in accessing very small time frames such as in Elliott wave theory and in MIDAS.
Ermanometry is based on the idea that financial markets exhibit multiple and concurrent orderly growth patterns in price and time, with time being the dominant influence. A given growth pattern is founded upon a ratio of expansion that never changes; this can be found in virtue of the methodology outlined under the subhead “Applying Ermanometry.” This ratio, or growth pattern, can be understood in terms of the log spiral, whose relevant characteristic is that its proportionality remains constant, whether it is increasing or decreasing in size.
Log spirals are common in nature, with one of the best-known examples being the nautilus shell. A fixed ratio (growth pattern) is also a feature of rectangular spirals, whose parameters are coincident with their log spiral counterparts. Consequently, Ermanometry shows that many important market moves share the properties of rectangular spirals, whose linear properties can be exploited mathematically to identify important market turning points in the constant ratio of expansion.
A central theme of Erman’s article is to show that the ratio influencing two important market moves between 1974 and 1978 has influenced the growth pattern of many subsequent moves up to 1998 in both the Dow Jones Industrial Average (Djia) and Standard & Poor’s 500. This is shown by taking various geometrical relationships from the rectangular spiral generated as a result of the growth ratio determined by the moves (seed segments) between 1974 and 1978 and then projecting them onto the market. This technique will be applied intraday under the next subhead.
Figure 2 illustrates the relationship between the constant growth of a log spiral highlighted in red and a linear rectangular spiral consisting of points A to J. The various geometrical relationships involve mathematical operations such as addition in relation to the segments and axes in the rectangular spiral. (Axes are the vertical and horizontal segments. For example, the axis EG is the hypotenuse of the triangle EFG consisting of two additional segments, EF and FG.)
FIGURE 2: GROWTH OF LOG SPIRAL (RED) AND LINEAR RECTANGULAR SPIRAL (POINTS A TO J). The geometrical relationships involve various mathematical operations such as addition in relation to the various segments and axes in the rectangular spiral.
Erman comments further on the nature of the rectangular spiral:
…Extension of any line in the series A-J, past the point at which it meets an adjacent line in the series to the point at which it intersects another line in the series, creates a rectangle. For example, if line GF is extended to intersect line IJ at point Z, then points Z, I, H, and G are the four corners of a rectangle. The axes segments have the same constant growth ratio as the spirals. Therefore, the diagram could be turned inside out using the axes segments as the rectangular spiral. The values for the previous rectangular spiral segments would then become axes segments. The shape of the log spiral would remain constant.
APPLYING ERMANOMETRY
With this introduction, let us look at the technique involved in establishing the fixed ratio (growth pattern) underlying a given price move. The first thing to do is to identify the start of a new trend whose growth pattern will be of interest. Figure 3 is a 60-minute chart of German Bund December 2010 futures with line (1) highlighting the first downside move off a significant swing high at the black arrow and line (2) highlighting the correction before the downside resumes. For the moment, we will ignore the dotted line. As we can see, the signals capture many important turning points at the green and blue pivots, especially where they cluster.
FIGURE 3: 60-MINUTE CHART OF GERMAN BUND DECEMBER 2010 FUTURES. Here you see a rectangular spiral launched from the red arrow.
The first downtrend segment (1) consists of 48 bars, while the corrective segment (2) consists of 40 bars. We use these two numbers — the seed segments EF and DE in Figure 2, respectively — to find the ratio:
48/40 = 1.2
The inverse ratio must also be found so that the spiral may wind counterclockwise as well as clockwise:
1/1.2 = 0.8333
From these ratios, we can derive the remainder of the segments on the rectangular spiral:
AB = 23
BC = 28
CD = 33
DE = 40 (seed segment)
EF = 48 (seed segment)
FG = 58
GH = 69
HI = 83
IJ = 99
We could continue increasing the segments, but IJ is the largest segment included in Erman’s article, and adding more would result in exceedingly large time gaps. Fortunately, all of this plus the calculations here can be programmed into MetaStock. Again, all a trader needs to do is input the time and number of bars of the two seed segments. Before moving on to the calculations, however, we need to make several important observations:
- Erman occasionally varies the derivation of the seed segments. In his article, his first variation involved seed segment (2) consisting of seed segment (1) plus the original seed segment (2) (see dotted line in Figure 3). In this case, the result is 48 bars and 88 bars, producing a ratio of 1.8333 and an inverse ratio of 0.5455. Seed segment (1) would become EF and seed segment (2) FG and the calculations would adjust accordingly. Erman will also sometimes take the square root of a given growth ratio (for example, √1.2 = 1.0954) and then recalculate the segments. Space considerations preclude me from experimenting with these options.
- As regards the launch point, Erman writes that “all seed points are valid points from which to project forward. … When a projected point does in fact become an important pivot, the new pivot is considered part of the spiral family and becomes a valid point from which to project forward.” In Figure 3 I started the projection from the end of the second seed segment at the red arrow. It results in a cluster of signals at the first green circle and another cluster at the second blue circle, marking the first pullback in the start of an uptrend that lasts several days (the actual bottom is three bars away from a signal).
- It is interesting to note that the time space between signals is never the same (the way it is in the Fibonacci and Lucas series), since the time space between each signal depends on the ratio. For example, the ratio 1.8333 is 30% larger than the ratio 1.2 and causes the spiral to expand and decrease more quickly. We can see the clustering of many signals in Figure 3. The same clusters will not be present on other charts as a result of the difference in the ratio.
- Ermanometry allows a variance of +2/-2 bars either side of a signal.
Suggested Books and Courses About Day Trading
With the segments calculated and various axes derived from them using Pythagoras’s theorem, we can now apply the spiral to this trend using Erman’s techniques, though it is important to bear in mind that he chose them because he was revealing how they were linked to the price movements between 1974 and 1978 (his two seed segments) and 1998. For the purpose of this article, I am going to add a few more techniques. Over the years, Erman has researched dozens of sophisticated geometrical techniques, and readers should consult his book Ermanometry via his website. Again, all of these calculations can be programmed into MetaStock.
FIGURE 4: ERMAN’S TECHNIQUES PLUS COLES’S ADDITIONS. Here you see Erman’s 1974 and 1978 seed segments plus a few added by Coles.
The two lists in Figure 4 contain Erman’s techniques first and then ones I have added. More permutations could be included, but these are more than enough for this article. Figure 5 is another 60-minute chart of the Bund moved forward, this time with the spiral launched from the start of segment (1) (red arrow) instead of the end of segment (2). The cluster of signals in the rectangle is redundant, since it is falling within segment (2). Green circles emphasize the main swing low and high separated by multiday price moves. The green arrow on the right indicates a projected point that now becomes a part of the spiral family, and hence a new point from which to launch a new spiral.
FIGURE 5: SAME 60-MINUTE CHART OF GERMAN BUND DECEMBER 2010 FUTURES MOVED FORWARD. Here the spiral is launched from the start of segment 1 instead of segment 2. The pivot at the green arrow is part of this spiral family and hence a new place from which to launch another spiral.
One concern I have with rectangular spirals is that several variables are associated with their launch, including:
- The choice of the two seed segments
- The actual launch point
- Whether the growth ratio should be further adjusted.
This raises the additional concern over whether the correct value for each variable has been chosen, an issue that is much easier to resolve when fitting rectangular spirals retrospectively as in Erman’s article. I address this by confirming a spiral’s ability to catch early subsequent pivots before making a commitment to it. For example, Figure 6 is a five-minute chart of the DAX December 2010 futures with a spiral launched from the red arrow using the seed segments highlighted. It catches three turns before the vertical black line, thus confirming that the values of the variables chosen are valid. A five-minute chart would not be chosen over a higher time frame chart for the reasons discussed earlier; I merely use this chart to illustrate that rectangular spirals do work on time frames smaller than 15 minutes.
FIGURE 6: FIVE-MINUTE CHART OF THE GERMAN DAX DECEMBER 2010 FUTURES. On this chart you see the confirmation of a rectangular spiral by the time price reaches the vertical gray line.
AN EFFECTIVE COMBINATION
There is no question that time-based systems are effective when applied intraday, provided they are automated to remove the need to calculate them manually. They can easily be combined, as Greenblatt has done with Lucas and Fibonacci, and as it is possible to do with DeMark and Ermanometry. However, this conclusion comes with two major provisos.
First, a time-based system should never be used in isolation from price-based technical analysis. The role of the latter is to provide a clear perception of current trend direction plus its strength (or tendency to break out if the market is consolidating). Critically, a price-based focus is also the only way of identifying setups and the only means of confirming them. Candlesticks (or an equally robust plotting style) are required for the latter at a bare minimum, with additional factors such as trendline analysis and moving averages also an essential means of confirming setups. Daytraders can also apply techniques such as floor trader pivots and Market Profile. Price-based analysis is particularly important in light of the fact that time-based signals will always be trend-following or contrarian and it is vital to know which side of the market you are on.
Second, time time-based systems of the type we have been surveying here are easily displayed on higher time frame intraday charts in relation to the trading chart time frame (for example, 60 minutes versus five minutes or 15 minutes versus one minute). Otherwise, the two issues discussed here may create some concern for traders.
Andrew Coles has master’s and doctoral degrees in the history of science. He has a diploma in technical analysis from STA-UK and from the International Federation of Technical Analysts (IFTA). He is also a Certified Financial Technician (CFTe). With his colleague David Hawkins, he has written a book on the MIDAS system for Bloomberg Press/John Wiley.
