Trend Determination

Trend determination is obviously important for trend followers, but anyone trading in any style should be aware of the trend in the market they are trading. The overall trend can influence your trading style. If the trend is up, you probably will have a different way you treat buy signals from counter trend sell signals. If the trend is sideways, then applying a trend following method would be frustrating and probably not profitable. A downtrend in certain markets can have a different character than an uptrend. The beginning of a trend can be easier to trade than end of a trend in many cases. Therefore, it is important to know what the trend is.

But to complicate matters, there can be many trends at play in the same market, even on the same chart. There are trends within trends. Different period lengths on a moving averages, or indicator input, can signal a different, often confusing and conflicting trend. There can be a counter trend down move on a 30-minute chart, while the daily chart is showing a powerful uptrend, while the monthly chart is showing a sideways trend. When they all line up, it is the most comfortable and reassuring time to take a trade, but if you wait for everything to be in synch you would probably trade very little, and often not in a timely manner. And often the comfortable and easy trade is the one everyone sees, and it often turns out to be untimely. My thought, after changing my mind on this issue many times over the years, is to keep it simple and just trade off the time frame of the chart you are analyzing.

And a further complication, if you are a trend trader, is that trending periods occur less frequently than non-trending periods. However, usually in non-trending periods tradable trends occur in the shorter time frames. And the longer, trendless time frame gives good reference points for the shorter time frame. If a shorter time frame isn’t possible to trade there are always markets somewhere in a trending mode.

Most of us want indicators to guide us, as indicators are quantifiable. We can lean on them with more confidence. However, the purest and fastest way to determine trend is just through studying the price structure. Price does not lag. Price is not derived from anything. It is current. It might frustrate, but it doesn’t lie.

tr1.gif
The above chart shows a series of price bars. I took the time frame and market off the chart for this illustration. It really doesn’t matter what time frame is being analyzed. A trend has the same basic structure in all time frames. The red and green horizontal dotted lines are placed at swing points in the price action. In this case I defined a swing point as any bar that has three lower bars, in the case of a high swing point, on either side of the high bar. There is necessarily a three bar wait until the line is formed. I don’t regard this as lag, since if the market kept advancing then a high point would not be formed. Some traders set the swing point only requiring two bars on either side. Some traders prefer to have it asymmetric by having two on one side and one on the other, or any combination you can think of. A low swing point will be the inverse, with a low bar and three higher bars on either side. This swing point is often referred to as a pivot, which is not to be confused with the floor trader pivots, which are not related. Bill Williams has done much work in the area of chaos theory, and he calls these swing points fractals. They are all the same. Most traders just eyeball an obvious swing point and draw a line extending to the right.

However you define the swing point, the idea is to have a series of prices that keep taking out the previous swing point in one direction. This is the same theory of a market making higher highs and higher lows, although this isn’t always exactly true. Sometimes a pause will form in the price structure and prices temporarily go opposite to the trend with the market making a lower high, but as long as a lower low is not taken out the uptrend is still intact. On the above chart there is a fairly smooth progression of swing points taken out from the first blue up arrow to the red down arrow on the right. There was only one instance, near the middle of the chart, where there was a slightly lower swing high.

I believe swing points are valid as a breakout strategy because the market views these points as areas to test. If the test fails the market will test or create a swing point in the opposite direction. If the test is successful the market will proceed upward, and the old swing point becomes an area of support if the market tests back down. The market is constantly testing whether trade is accepting or rejecting price, and these swing points are the reference points for these tests. Don’t trade without them. And don’t trade off an indicator without prices.

A quick word about Elliot Wave (and it is only my opinion). Many traders try to keep track of these swing points by counting waves. There are often five waves in one direction, three impulses and two reactions, and after this sequence there will be three waves in the opposite direction. There are many variations on the corrective phase, or the three waves in the opposite direction. That’s part of the problem. There are so many variations, and there are always the alternate counts that occur when the wave forecast is incorrect. It is easy to see the wave structure in hindsight, but very difficult in real time. Rarely will two Ellioticians agree on a count. The real problem with Elliot Wave for me is getting a bias on something that you think should happen, based on where you interpret where you are in the count, and then expecting it to happen, rather than letting the market tell you what it wants to do.

tr4.gif
Perhaps the most common indicator for determining trend is the moving average, or in the case of this chart, the dual moving average. You can see at a quick glance if the faster moving average is above or below the slower, so trend determination is quick and easy. However, there is often much lag. This chart was a well chosen example of an easy trend. Most times moving averages make many false crossovers just at the wrong time.

tr3.gif
The above chart shows the same price series with a linear regression curve along with standard error bands. There can be considerable lag with this method as well. I will discuss this indicator in much detail in the indicator section. But briefly, I find it better to trade only on the side of the regression line (the dark solid line). The timing signals can be taken from another indicator. I’m only interested in trend direction to determine which side of the market to trade. The error bands give me additional information. When they are wide the market tends to be noisy and not trending well. When the bands start to narrow, the market it starting to trend more smoothly. The bands often offer good support and resistance levels. During trend changes you often see prices hugging the regression line as it changes direction, and the bands tend to contain price. The breakout of this containment is often a good entry point. A perfect example is on the left side of this chart.

tr5.gif
The above chart shows swing points, but instead of a well chosen chart with a nice uptrend, this chart shows a more difficult environment to trade. There was a nice series of swing point highs taken out on the left side, but then the market became less agreeable. Swing points began to be taken out, but with no follow through and no direction. But just when most traders would want to give up in frustration, on the right side of the chart a trend started to emerge.

tr8.gif
The above chart shows the same price series with the same moving average as in the earlier trending example. If you try trading the crossovers you usually give back more in the chop than you make in the trend.

tr13.gif
Here is the same price series again. This time with the linear regression curve and standard error bands. It isn’t perfect, but it does show the trend tendency even through the choppy area. If you only traded the short side through the chop you at least wouldn’t get chopped back and forth as much, and would generally be on the correct side of the market.

tr6.gif
How do you tell in advance if there will be trend continuation, or the market will turn to chop? You can’t. There are tools to help guide you, though. One method to help determine the quality of a trend is Welles Wilder’s ADX. It is perhaps the most common indicator to try to determine when the market is trending well, or not trending. I have never been successful with the ADX. I point it out because it’s the most common indicator used for this. I much prefer the R-Squared. I have an article on this indicator in with much more information in the indicator section elsewhere on this blog. It is a correlation study. In this chart it measures the accuracy of prices in relation to a linear regression curve. There are many other applications for the R-Squared, as you can plug in any two variables and compare their correlation. When the red line on the indicator is above the blue line, it shows a persistent trend. Little dots are on the line when it has been above the blue line for a certain number of bars. When the red line is under the yellow line, prices are displaying little or no trendiness. You can see on the chart, when the trend was in place on the left side, the R-Squared indicator trended up with prices, and went over the blue line. When the market went into a choppy, sideways mode, the indicator retreated and quickly went under the yellow reference line and mostly stayed there throughout the chop. On the very right side when the market started to trend again, the indicator came out of the low level and showed trendiness. There will always be lag in this type of indicator, but nowhere near as much lag as with the ADX, at least in my opinion.

tr7.gif
Here are the same swing points along with the R-Squared indicator. This time I’m showing a daily chart of the S&P Energy Spider, current through the date this was written. This was a very nice trend, with the R-Squared firmly above the upper blue reference line. Sometimes there is not sufficient warning when a correction will come, as the line often stays above the blue line, indicating trendiness, well into a correction. There was one point near the right side of the chart when prices took out a low swing point just as the R-Squared was crossing the blue line. Momentum divergences, along with a declining R-Squared, even though still above the blue line, can sometimes give a clue. There are no perfect indicators. But there are many tools that can help at certain times. I’ll have more about trend exhaustion and trend change in the next article.

tr11.gif
Here is the same energy spider with the linear regression curve and error bands. It catches the trend nicely, offers good support on pullbacks, and narrows slightly showing that the trend is smooth, despite the little sell off a few bars back.

tr16.gif
Here is a chart of Freeport-McMoRan with the same regression curve with error bands, but here I added the adaptive CCI (for explanation of the CCI, there are three articles in the indicator section). You can see when the histogram bars on the CCI are green and above the zero line, which is the dark purple line, that the uptrend is intact. Pullbacks to the zero line can be bought as long as prices are above the linear regression curve, and even better if they are above the upper error band (the dashed line).

I should mention another trend following approach that was very popular in the past. It is the channel breakout. I did not include a chart, as I don’t think it to be a viable system in today’s markets. It worked in a very narrow window of time when commodity markets trended well. The basis of this approach is to buy the breakout of the high of x number of bars ago, or short the breakout to the downside of the low of x number of bars ago. Often a stop is used with the same breakout method, but with a shorter lookback period. One of the originators of this method, who was incredibly successful for a time, has not been able to be profitable with it in recent markets, even after many attempts. I think a more valid approach using breakouts is to trade the swing points, rather than a fixed lookback. The swing points are dynamic and reflect the structure of the price action. Fixed lookback periods don’t offer a breakout related to market structure. The market is testing swing points to see if price can be accepted or rejected at these points. The market doesn’t care about testing a price bar that occurred 20 bars ago just because it was 20 bars ago.

I have not found any trend following indicators or approaches to be profitable as a stand alone system. But I have found it not profitable trading against the trend. Therefore I want to know the most probable direction of the current trend of the market and time frame I am trading. Then I can enter on pullbacks against the trend, but always in the direction of the trend. It’s taken me many years of frustration to learn this lesson. It might be gratifying to pick a top or bottom of a market, but there will most likely be higher odds of success in going with the flow of the trend. If you jump into a raging river you can either flow with the current effortlessly, or swim like mad trying to go upstream. Usually the best you can do is stay where you are.

The next article on trend will be ways to find trend exhaustion and to know when the trend has changed direction.

4 thoughts on “Trend Determination

  1. I like your analysis. More powers to your elbows. Please am a newbie in Forex. o, can you help me with a good system that trades 5 to 30 minutes with common indicators.
    I will be happy to get it fro you.

  2. Leave chop at bottom and at top sign is 40 ema is almost horizontal.Our grandmothers in India while milking a cow leave few streams of milk for mother Earth and last few streams for child of cattle.Middle of trend is pure milk leave bottom and top to other guys.

Comments are closed.