Back to top

Image: Bigstock

Trend Lines Are Mathematically Absurd

Read MoreHide Full Article

The most important lesson of the 2008 financial crisis for Wall Street was that risk models like VAR (value-at-risk) were fundamentally flawed and created a situation akin to children playing with matches -- in a fireworks factory.

They were flawed because they were built on the simple mathematical construct of standard deviation, which was designed to measure the variance in physical domains, like animal height and weight, and populations of data with predictable patterns and margins of error, like test scores or Brexit polls.

That Great Intellectual Fraud

The bell curve “normal” distribution takes data and crunches it down to handy numbers so we can talk about averages and “deviations” from normal. But it was a disaster for Wall Street to use it as the foundation of risk measurement in complex portfolios of mortgage-backed securities (MBS), like the CDOs (collateralized debt obligations) that took down Bear Stearns.

I talked about these themes in my recent video-blog on the film The Big Short. I also gave credit where it was due to the man who predicted the financial firestorm, Nassim Nicholas Taleb.

Taleb’s seminal book The Black Swan brilliantly laid waste to standard deviation, VAR, and the bell curve -- “that great intellectual fraud” – as purveyors of mystery, stupidity, and the arrogant certainty that produces financial weapons of mass destruction.

My Target: Drawing Straight Lines on Curvy Models

We owe it to Taleb for giving us major insight into the limits of financial models in forecasting and managing the “wild randomness” of markets as social beasts with infinite complexity.

I am taking the strength of his argument to now kill a much smaller, but no less sacred, cow of markets and trading: the trend line. In the school of technical analysis, a trend line is a straight line drawn to connect price extremes and thereby display a trend direction, as well as a visual marker for its violation.

In the video that accompanies this article, I show several examples of trend lines. I also compare and contrast the use of trend lines with something they seem similar to: linear regression lines which are commonly used for predictive analysis with a scatter plot of data.

The #1 Reason Trend Lines Are Mathematically Absurd

I also explain in the video why this tradition of drawing lines on charts is old and venerable, and why it will annoy people that I am calling them “absurd.” In 1948, Robert Edwards and John Magee published their book Technical Analysis of Stock Trends.

In its 10th edition, it is considered the “bible” of the chartist community and it still heavily promotes the use of straight lines connecting price extremes across time frames of one century to one minute. No one questions the validity of the extremely wide-spread practice. So I am really picking on a holy sacrament of the church of charting.

The reason trend lines are mathematically absurd is that markets are not linear. In fact, as Taleb taught us, markets are the most complex thermodynamic systems on the planet with far too many variables to allow any model to have much better than a coin flip’s chance of consistently predicting outcomes and volatility in any time frame.

In essence, market prediction makes weather prediction look like child’s play, and we know how we love to gang up on the weather forecaster who gets the high wrong by 5 degrees, 5 days out.

The question that chartists and others who draw trend lines have to ask themselves is this…

Why would an extremely complex behavioral system obey an arbitrary straight line through time?

My video has all the charts and slides to fully illustrate this fascinating and important discussion. Now put down that ruler and watch it!

Kevin Cook is a Senior Stock Strategist for where he runs the Tactical Trader portfolio.