Trend Lines Are Mathematically Absurd
In January of 2016, I wrote a short post on StockTwits to point out a gigantic fallacy in the world of trading stocks, securities and other assets using price charts. That world is commonly known as “technical analysis.”
I come from that world, so I felt sufficiently informed, experienced, and entitled to opine on this fallacy. In fact, I partially believed in at least two of the core tenets of the technical analysis of price, which often ignores company or asset fundamentals:
(1) asset fundamentals are to a large degree reflected in price as investor-trader beliefs about those fundamentals show up in the buying and selling behavior of many diverse participants.
(2) the future price of an asset, across different time frames, could be forecast and bet on using Bayesian probability by analyzing price patterns, mathematically-based momentum indicators, and historical-behavioral tendencies.
In extreme price moves, across and relative to different time frames, the greed and fear of participants stretches and bends fundamental views (beliefs) and provides additional opportunity for the astute “chartist.”
Overall, hundreds of mathematically-based technical indicators and historical patterns can be statistically constructed, tested and validated across multiple time frames to aid in forecasting the direction of price.
But not for this particular non-statistical, price-derived fallacy —the trend line — that is akin to using numerology for forecasting the winning Powerball ticket.
The title of my January 2016 post was “Trend Lines Are Mathematically Absurd: Killing the sacred cow of technical analysis in 400 words.”
While I was expecting all kinds of reaction, both from devout trend line practitioners and new converts to my thesis, I received very little response. I believed I was exposing the lunacy of a particular sacrament of the holy scriptures of charting, one based on ignorance of statistically complex systems. And one preserved by blind belief in tradition.
Where I was looking for incredulous disdain for my idea and analysis, I was met with indifference, brief brush-offs, and shrugs. And still, the professional community of technical analysts went on using trend lines, believing they are meaningful, discussing each others’ straight lines, praising each others’ insight, perpetuating each others’ faith in what I call “zodiac finance.”
So in June of 2016, I decided to make a video to detail the thesis, with lots of pictures and a lot more than 400 words. Still, little reaction; 18 months later, that video only has 1,000 views.
I have embedded the video below. But let me just explain why the argument might need more than 400 words: because there are many ways to talk about the usefulness of multiple trend indicators without committing the intellectual sin of drawing lines with rulers across the millions of multidimensional, social variables and billions of data points that create price.
Here are those 400 words from January of 2016, and then I will explain why trend lines are beyond mathematically absurd and border on being a sin…
The reaction of many chartists has been and will continue to be, “But trend lines work for me. What’s the big deal? Where’s the sin?”
As someone who takes very seriously the education of youth, here is the sin:
Mathematical and statistical knowledge and skills are hard enough for youth to earn, that we should be very careful about using so-called tools that pretend to be knowledge-based and finance-based but serve only to confuse, or purport magical thinking.
There are comparable examples of pseudoscience throughout history, like phrenology, astrology, numerology, and the bloodletting that killed George Washington.
And still the chartist might say, “But most tools in technical analysis are mathematically-based, like moving averages, stochastic oscillators, and Fibonacci retracements. What’s wrong with this one transgression that I find useful in my art, even if it’s just a tradition that everyone else is doing?”
My answer: by preserving this one, widely-popular pseudo-math analysis, you discredit the profession of technical analysis and teach youth that charting the markets is akin to the zodiac, completely subjective, non-quantitative, and magical.
Here is a subsequent useful comment I added below the video on YouTube…
“Another thing I left out of the discussion was that one can still see and draw patterns on charts, like symmetrical triangles. You just don’t need (and shouldn’t use) “ruler-straight” lines that pretend there is precision extrapolation going on. All you need is a rough line to see the pattern.”
The reason I grant this distinction — for smaller time frames of 1-hour to 1-year — is because it often helps to see how price is consolidating as collective behavior absorbs new levels and information. The historical patterns that chartists have come to know and love still work as visual guides to buying and selling pressure. We just don’t need anything more precise than can be drawn by hand or with a software pencil.
And pretending there is simple precision here is what causes the most anti-science damage. You’ll see this fraud really stand out when a “ruler junky” draws a trend line on a multi-year chart of the stock market that downplays, if not ignores, the billions of economic actions, and trillions of dollars, between two price points connected by a straight line.
Without further ado, here is my presentation that you should watch, comment on, and share to expand the conversation. And I am always open to someone using math to prove me wrong…
Thanks for watching!
Kevin
