Any pair of intersecting trendlines contains information that enables the forecasting of major price inflections (see “Spotting momentum swings in advance,” November 2009). This concept has even greater relevance when applied to index exchange-traded funds (ETFs) and exchange-traded notes, and is further enhanced by the introduction of synthetic trendlines. Here, we’ll offer an easier method that can be used to measure the distance from the trendline intersect, simply by counting the candles.
The nature of any index is either that it represents a basket of issues rather than any single issue, or it represents a divergent group of market participants or points of sale. This makes it less vulnerable to minor news and emotions affecting any one issue. Surprise market shocks, and the unknown results that follow them, are dampened. This is why many traders prefer index trading to trading any one issue or commodity. It also makes indexes more answerable to mathematical forecasts.
The technique Forward-Looking Information Radar (FLIR), as described in the November 2009 article, requires several complex steps. For those steps, see the earlier article. However, a simplified approach is to measure from the intersect of any two trendlines to any inflection to the left of the intersect, and then measure that same distance out in the future to the right of the intersect. That point almost always is an inflection in the price trend (see “Creating FLIR,” below). FLIR is surprisingly accurate, seeming to owe its precision to a geometric integration of time and price.
FLIR does not disclose the likely directions of trends on either side of the future inflection point. It also requires precise measuring and placement of trendlines, without which error may occur. To enhance FLIR’s effectiveness and to avoid these problems, the trader can add synthetic trendlines to the analytical mix.
Synthetic trendlines come as two pairs of geometrically related trendlines. They are simply manual graphical drawings that are constructed based on basic spatial relationships taken from price action and then transferred back to the price chart.
Synthetic trendlines should be thought of as probability pairs, rather than definitive statements. Price almost always will follow one of the lines, though it may take two to four bars before choosing one. The intersects also define inflections. Further, synthetics geometrically incorporate rate of change in price over time, as the integral of two previous trends.
Three reference inflections are selected, forming a V or inverted V (see “Creating synthetics,” below). The vertical distances from the vertex of the V to each of the reference points define the lengths of the sides of a rectangle. Lines connecting the corners of the rectangle integrate the price differentials, creating one acute and one oblique pair of angles. Placing these angles at the reference points integrates time differentials — thus, a future price path is proposed.
Obviously, placing each of the four pairs at three different points results in 12 possible paths, but that much complexity is illusory because some paths will overlap and others will seem outrageous, even in a volatile market. After a few practice applications, you’ll find yourself placing only two or three pairs, and within a few periods this quickly narrows real pathways to only a couple of possibilities, then just one.
In dynamic plots (indicators) using pure math, time and price typically are integrated by means of moving averages, and are of the nature of real-time indicators. Synthetic trendlines sacrifice real-time behavior in favor of future price behavior. They extract from time and price a theory of where price could go, giving us idealized paths. The sacrifice must be made because, within a single function, we cannot know direction and magnitude simultaneously. (Traders are in good company. Quantum physicists have the same problem concerning position and momentum of high-energy particles.)
FLIR produces a one-dimensional event narrative; that is, at point X, the price path will reverse. Synthetics generate a two-dimensional event narrative, more complex and probability-based; that is, from point X forward, the price path will be one of these lines, and will reverse at this later point Y. Combining the two can yield a powerful inflection picture as well as the duration and magnitude of future moves.
Understanding the synergies
But why are FLIR and synthetics so successful? The answer seems to lie in the regularity, or memory, of markets. Old trendlines never die; they just fade in and out as support and resistance. So they remain a factor in many indicator studies for long periods, though the math connection may not be obvious to the user. What also may not be obvious is that trendline divergence often is incorporated into many indicators, which helps explain FLIR’s success.
For example, moving average convergence-divergence (MACD) contemplates all price within a period, including inflections. Another way of saying this is that each moving average in the MACD pair is a limit function with one high and one low parameter, and intermediate values moving toward these parameters.
At some point, a price inflection occurs. At this same point, there is a specific amount of divergence between the two moving averages, shown by the histogram. The next time that amount of divergence occurs, large market participants (or their high-speed algorithms) remember it and behave accordingly, creating an inflection. This divergence value and market response may be repeated several times before seeming to expire, and new values are added all the time. That divergence is represented graphically in FLIR. An explanation for the success of synthetics is similar, though we’d have to add a trendline to the MACD itself, a method far more cumbersome and far less visually informative than synthetic trendlines.
In lengthy trends, FLIR is at a disadvantage because it’s hard to get a good pair of trendlines to intersect in an accurate and useful way. By also using synthetics, we can identify a great many more inflections. FLIR accurately locates inflections about 85% of the time, allowing for one-period slippage. Failures range from simply no inflection to the rare and disastrous case of mistakenly flagging as an inflection what turns out to be an acceleration in the current trend. Synthetics just indicate possible trends and, to date, haven’t been known to indicate an inflection that turned out to be an accelerating trend.
So why use FLIR at all? Why not just use synthetic trendlines?
The simple answer is that FLIR is easier and faster to use than synthetics, though with a little practice you can generate synthetic trendlines in less than a few minutes. FLIR also has a better likelihood of noting large inflections if you create trendlines using major tops and bottoms, as shown. Synthetics, though giving you an abundance of inflections of all sizes, are far more useful in providing a clear idea of trends, support and resistance.
To illustrate these strengths and weaknesses, let’s walk through some actual examples. We first have “Trending corn” (below), using only synthetic trendlines (the steady uptrend would have made FLIR difficult). We get clear inflections — some major, some minor. Trendlines, support and resistance are shown strongly across a future span of nearly 11 months from the date when we first could have made the forecast.
However, we can apply one trick for using FLIR on a difficult chart, such as the corn example. In “FLIR fix for corn” (below), we perform a small feat of mathematical prestidigitation by using the horizontal distances between inflections instead of the height between trendlines. This works just fine because the math is embedded in the algorithm.
Next, we have two identical Dow Jones Industrial Average charts, the first showing FLIR capturing strong inflections. The second, a synthetic trendlines chart, begins by giving us a clear uptrend line, followed by clear support and resistance levels, along with large and small inflections (see “Index trends,” below).
When applying the indicators, FLIR is simple to do, requiring little skill, and can be done directly on most charting programs. Distances can be measured by counting candles. Synthetics, on the other hand, require some modest skill with a drawing program. Depending on the charting program you use, you may have a capable drawing program built in.
In setting your synthetics, use both the acute and oblique angle pairs, and test their placement on each of the three anchors for relevance; if irrelevant, delete to avoid clutter. Using different colors for obliques and acutes also is helpful. The best short-term synthetic uses the most recent triad of anchors. It’s also perfectly fine to skip inflections, especially if they’re small, but bear in mind that older synthetics expire, replaced by newer ones.
By the time you’ve created a dozen sets of synthetics, it will seem easy. And it’s worth it. By combining FLIR and synthetic trendlines — or by alternating them according to need and circumstance — the trader can pry out the desired information for profitable entry and exit on indexes and their ETFs.
Dennis Hudson is a quantitative trader and former General Electric diamonds engineering researcher where, in the 1960s, he used his quant skills to determine particle size distributions. He may be reached through his website, www.orangequant.com.