Trending or Not? (Part One)

Analyzing or trading a market can often be a very frustrating experience, and most realize that the frustration makes trading harder from a psychological standpoint. One of the more frustrating situations for me as an analyst is when I come up with a plan and execute the entry correctly, only to exit midway through a strong trend. Another is when I am expecting a market to start a dramatic move, but do not have a position when it starts, and the initial move is so strong that the new entry risk is too high. This leaves me to watch the action from the sidelines. Both of these can have negative psychological effects, and for me, if taken out mid-way in a trend, it was often difficult to get back in. I am sure many of you have had similar experiences and frustrations, regardless of what time frame you trade. Over the years, I have often tried to come up with a method or indicator that would allow me to stay with a trend longer or give me a high-confidence sign that the market was trending so that a higher-risk entry level was acceptable. In this series of articles, I will review some of items I have looked at in hopes that they may assist you in your trading.

One of the many strong trends that occurred during 2008 was in the heating oil market, which was significantly more stable than the crude oil market. The chart below uses the continuous contract weekly data and includes a number of indicators I would like to review before continuing. One indicator that I have used for over twenty years is the ADX, which I use along with a 21-period weighted moving average (21-WMA) to help identify trending and non-trending environments. The best signals I have found come when you have an ADX that has been declining for 40 or more bars and dropped below the 20 level. Often times there are two peaks within the decline, which allows you to draw a nice downtrend. During most of the ADX’s decline it should stay below its declining WMA. The signal that the market is moving from a non-trending environment to trending comes when the ADX moves back above its WMA and breaks its downtrend.

Most of you are also familiar with Bollinger Bands and the chart below has the standard inputs, 20 periods with two standard deviations. On the bottom of the chart I have plotted the bandwidth (BW), which is the difference between the upper and lower bands. When this gap is small, volatility is low, and when they start expanding, volatility is increasing. I have also plotted a 21-WMA of the BW. For this first chart, a market will be considered trending when the BW stays above it is WMA for three consecutive periods. Conversely, three consecutive closes below its WMA means it is non-trending. I generally prefer methods of analysis where two indicators that have somewhat different derivations are in agreement, so I will use the following process in examining these two methods. In order for a market to be trending, the ADX needs to be above its WMA and the BW needs to have been above its WMA for three consecutive periods. A non-trending (NT) market is when the ADX is below its WMA, and the BW has been below its WMA, for three consecutive periods.

Figure 1

The weekly heating oil chart covers the period from 2006 through early 2009. The ADX line declined for most of 2006 and reached a low of 14 in the middle of June before rising above its WMA in early July, with heating oil closing at $2.04. A long-term trending signal was generated the week of September 23rd when the long-term resistance in the ADX, (line 1) was overcome (point a). Heating oil closed this week at $2.25. It was not until early November (line 2) that the BW moved into the trending mode with heating oil’s close at $2.57. Now, I’d like to note that I do not advocate buying or selling just because a market is trending or not. Rather, I use this information to adjust my risk management. In a trending market, I give the position more room, but if it changes to non-trending, then I recommend either reducing position size or tightening stops.

In the next three weeks, heating oil hit a high of $2.72 before a two-week setback. Over the next nine weeks, heating oil did make higher highs and higher lows, reaching $2.74 in early January 2008 before correcting more sharply taking the prices back to the $2.40 area. Two weeks after the lows, the market moved into non-trending mode (line 3). This was not a timely signal, as four weeks later heating oil completed its consolidation pattern by closing at new weekly highs. By early April (line 4) the market was trending once more, with heating oil closing just below $3.00 and well above its correction lows. For the next three months, heating oil had an impressive run, reaching a high of $4.15 before reversing the following week to close below the previous four-week lows, which was a bearish indication. One week later, the ADX dropped below its WMA (point b), and three weeks later, the BW also changed to non-trending (line 5). As heating oil prices collapsed with crude oil, the BW moved back to trending in early October with heating oil at $2.66. It took another 50 cent drop and five weeks for the ADX to move above its WMA (line 6). Heating oil closed the year at $1.27 with both indicators still in the trending mode.

Another market that had a strong uptrend in 2008 was corn, which was helped by both the ethanol craze and flooding in the growing region. For most of 2006, corn was non-trending as the BW was flat and at low levels. The week of October 29, 2006, the market started to trend (line 1), with corn closing at $3.32. Both indicators stayed in trending mode until late January 2007, when the ADX dropped below its WMA (point a), with corn closing at $4.05. The BW stayed trending until the week of February 17, 2007 (line 2) when corn closed at $4.17 and prices moved higher for another two weeks before topping out at $4.49. Therefore, the implementation of a profit-taking strategy or tighter trailing stop after the shift to non-trending would have worked well. For the next ten months, corn formed a long continuation pattern between resistance at $4.49 and support at $3.17, which was completed the week of December 8, 2007. The ADX line had reached a low of 12 during the fall, and at the start of December, it moved above its WMA (point b). The BW did not move into trending mode until the end of the month (line 3), with corn closing at $4.52. For the next seven months, there was always one of the two indicators in the trending mode as corn hit a high of $7.96. The BW dropped below its WMA from late April through mid-June, but then stayed trending through the peak in corn prices and the subsequent decline. The ADX dropped below its WMA the week of July 20th just after corn formed an island reversal and prices plunged.

Figure 3

After reviewing the weekly data, lets see if the daily analysis could be used to clarify or improve the weekly signals. When looking at the daily analysis, is it important to keep in mind that from December until early-July 2008, the weekly analysis indicated a trending market. The daily ADX line, which had been below 15 for most of August and September, broke out of its trading range (line 1) in early December. A few days later, the BW changed to trending (line 2) as corn closed at $4.12. The ADX lines switched to non-trending on January 17th (point a) as it dropped below its WMA with corn closing at $5.02. Six days later, the BW closed below its WMA for the third day with corn settling at $5.02 (line 3). By February 20th, the ADX was back above its WMA, and nine days later, the BW was also trending with corn prices 26 cents higher (line 4). This signal was reversed just two weeks later as the ADX dropped below its WMA on March 13th, with corn 20 cents higher. The BW became non-trending on March 28th (line 5) as corn had dropped 12.4% from the March 12th highs to the March 21st lows before regaining all this ground in the next six days. This sharp correction skewed the daily analysis, while the weekly remained clearly trending. Both the ADX and BW were back in the trending mode with the close on April 5th (line 6). This was reversed on April 23, though corn had only moved about ten cents higher. Corn was non-trending for the next six weeks as it traded in a 60 cent range. The breakout from this range moved both indicators back into the trending mode on June 7th. Corn rose from the June 7th close at $6.50 to an intra-day high of $7.96. The ADX dropped below its WMA, and the BW changed to non-trending with the July 3rd close at $7.80. Corn gapped lower two days later and began a vicious plunge that took prices back to the $3.00 area in December 2008.

Figure 4

From looking at the weekly and daily data on corn, it seems as though one can conclude that the ADX works better than the BW, especially on the weekly data. For example, the weekly ADX on corn stayed above its WMA from November 24, 2007 through July 20, 2008. A disciplined trader might have used swings in the daily ADX readings to scale in and out of a core position. Alternatively, a trader could also try a trailing stop method to keep in a position. To illustrate this possibility, I have added a stop method discussed in a previous article (see Average True Range). The formula for stops on long positions looks at 20-MA of the highest high over the past 20 periods and then subtracts three times the 14-period ATR to calculate the stop. When the weekly ADX completed its move into the trending mode on November 20th with corn closing at $3.89, the ATRLSma was at $3.61 (point 1). Corn did not close below the ATRLSma until the March 21, 2008 close at $5.21 (point 2). Corn prices, as noted earlier, reversed rapidly from these lows with a new closing high on April 2nd of $5.84. Despite an intervening high of $6.37 during mid-May, Corn again closed at $5.82 (point 3) on May 30th, below its ATRLSma. Just two days later, as corn closed above it’s prior swing high, the daily ADX broke its five-month downtrend (line a) and clearly had moved back into the trending mode (point 4). The stop held until the close on July 12th (point 5) at $7.04, as corn began its long decline, dropping close to $3.00 by the end of 2008. Of course, using this stop on weekly data using the same starting point would have just kept you in from approximately $3.89 to $5.18.

So is using the weekly trending signals and the daily exits a feasible strategy in today’s market? I believe it is for some traders, but not for many. Fifteen years ago, there were many traders who would just catch a move in one market (such as the one in corn) that would make their entire trading year. Technology and other factors have likely decreased the number of position traders significantly, however, as a long-term advocate of multiple time frames, staying consistent with the weekly trending analysis can keep you out of trouble by reducing the number of trades you take against the major trend. The ADX, in my estimation, is a very good tool for alerting one to a change from a non-trending to a trending mode, but does not do as well, based on the daily data, in keeping you in that trend for the long term. In the next article of this series, I will look at some other methods and indicators such as the DIosc to determine whether they can help you to stay in significant trends regardless of whether you trade based on weekly, daily, or intra-day data.