Trend Identification Algorithm

In the old days, as described in Investtech’s book on technical stock analysis, visual technical analysis was conducted manually: "Study price movements in order to identify a trend. The price rarely moves in a straight line, but rather as a series of tops and bottoms. If you can draw a straight line between two or more rising bottoms, this is the support line of a rising trend, see figure 1. Then draw a parallel line through the rising tops. This line is called the trend’s resistance line, and support and resistance combined make up the trend, see figure 2.

Similarly you get a falling trend if you can draw a straight line through two or more falling tops and a parallel line though falling bottoms

Many investors still identify trends by studying charts and drawing trend lines by hand. The key weakness of this method is its subjective nature, allowing you to see the trends you want to see which may also fit in with your subconscious preferences. It is also very time consuming.

Now, though, good tools for such analysis are available. Investtech has developed automatic algorithms for trend identification. Every day these find the ”best” trend in the chart, where included criteria amongst others are the density of the trend, deviation from regression and number of points near floor and ceiling. In an Investtech chart for the medium long term, with 18 months’ history, approximately 80,000 different trend alternatives are assessed every day. Each is assigned a score, and the one with the highest score is selected.

Figure 3: It can be hard to tell whether a stock is in a rising, sideways or falling trend, and whether a trend channel is broken out of or not. It is useful to have a tool that identifies this based on objective mathematical criteria.