Trend Analysis

A lot of people like to select stocks depending on their trends. There are even entire TV programs devoted to trend analysis, people predicting to what price target a stock may decrease or or increase. It is quite simple to demonstrate how easily we can be fooled by a so called trend. Let's start with the following test question: Please choose which of the following 3 stocks you think shows the most compelling signs of a trend:

x

x

x

 

ANSWER

While we could draw plenty of lines into the charts above to 'explicitly show' that there are plenty of trends in those stocks, the truth is that none of them has a trend. Let me explain you why:
All of the charts were produced by a computer program that generates a random walk. This means, the program moves the stock each 66% of the time between -1 and +1 with an average of 0. 99% of the time the random value it generates is between -2 and +2 (this corresponds to a so called normal distribution with an average of 0 and a standard deviation of 1).

By definition, for each graph there is a 50% chance at each point in time that at any point in the future the stock will be higher and a 50% that it will be lower.

The example demonstrates how easily our minds can be fooled through our constant quest to find patterns in objects, even if there is no pattern at all. For your own amusement, feel free to compare the charts above with any stock you like.

You may convince yourself with the following example: press the button to generate a random walk yourself!

 

 

 
 
   
     

Comments

Name


Email (optional, will not appear in the comment)

Comment



 

   

 

 

 

Hollie

That's way the bestest ansewr so far!

 

Armando Alizo

Dick, While I agree that you can generate charts that "look" like trending securities using random variables, that doesn't prove that Trend Following as such is an invalid strategy. There are many ways to define a trend - some may be profitable in some markets and not in others. Naturally ANY approach (trend based or not) must be properly formulated and back tested. I do agree completely with your comments regarding TV "gurus" that use charts as if they were tea leaves to predict the future. The future is unpredictable, and in my view it is actually COUNTERPRODUCTIVE to attempt to make forecasts. Such approaches are utterly subjective, inconsistent, and unprofitable.