Seeking Your Friend the Trend? Try Charting Linear Regression

Traders have been debating the best way to identify a chart trend since Charles Dow first assembled his famous indexes. To say the least, it can be a tricky process. But once you’ve embraced a method, the trend can become your friend.

Yet, trends—and friends—aren’t all created equal. Identifying peaks and troughs is a good start, but might leave some traders with only a surface relationship with a particular chart—not enough information to make quality trading judgments.

Some traders try to use Dow’s method of monitoring peaks and troughs in the price action. Figure 1 demonstrates how difficult it can be to tell which peaks and which troughs to use for your next move …  especially when to use them. The chart is moving higher, but some peaks weren’t creating the higher highs and higher lows needed to suggest the trend.

Other investors use a moving average to identify the trend. But which moving average is best? There are geometric moving averages, exponential moving averages, triangular moving averages, and many more. Then you have to determine the length: 50-day, 100-day, and 200-day moving averages are all common. Even when you’ve decided on a type and length for your moving average, it can still curve and bend so often that it’s hard to say if the trend is changing or not (figure 1). 

But I Regress…

One way to spot the trend is to allow the data to speak for itself through linear regression. Linear regression is the analysis of two separate variables to define a single relationship. On a stock chart, this is the relationship of price and time.

The linear regression trendline in figure 2 charts a mean or average line for all price points over a certain time frame. The slope of the line gives the trend. Once the trend is defined, a trader may opt to place trades in the same direction.  

Channeling the Trend

One way a trader could tap linear regression is to layer on a 100% linear regression channel to act as support and resistance (figure 3). The trader would presumably buy on a support bounce and sell at resistance. 

Some traders opt to reduce risk by placing a stop order on entry below the lower channel. 

Once the price reaches the first target, the stop order is changed to the entry price, a measure that is intended to reduce the trade’s risk-to-reward profile. Targets for the mean and the upper channel must be set on the day of the bounce, because the lines will change as new data appears and old data disappears. Even if the lines change, pre-set targets won’t.

 Of course, when trading a linear regression channel, most traders believe they should stick with the trend. And that’s why linear regression is one potential way to make the trend your friend all over again.