Friday, May 11, 2012

Don't be Fooled by Pattern Recognition

The Lure of Pattern Recognition (Be Very Careful)

Everyone who is in the market is looking to do the same thing, make money.  Institutions and hedge funds call this "alpha".  The game of Making Money is a very competitive environment.  When you place a trade you are competing with highly funded well staffed opponents.  The efficient market hypothesis says that you can't beat the market because all available information is already available and in the market (

Individual investors typically have zero access to advanced computer models so they rely on the one thing that is available to them for free, intuition.  They they look at charts, draw trend lines, and infer patterns to make investment decisions.  The problem is that there is a large amount of scientific evidence that suggests our brains are wired poorly for this type of visual analysis. Consider the case of eye witness testimony.

False Misidentification
Alarming statistics state that in 2000, of 62 cases that involved the conviction and imprisonment of individuals for felonies, 52 of them-including 8 of them that had death sentences-involved mistaken eyewitness identifications. These individuals were later acquitted of these charges with the use ofDNA testing.

Double Bottoms, Triangle Tops, Head and Shoulders, Support, Resistance, Waves, all of these terms are used to describe price action of a stock.  The problem is that when you see it, you are looking at an instance and then extrapolating a pattern.  You have no idea how many other stocks are doing the same thing and what is the actual post observational result of the price pattern.  If you had the data you would see that most likely you are not seeing any kind of pattern at all, but rather your eyes are just attracted to creating recognizable shapes.  It makes you feel comfortable and says, "Hey I know what is going on right now."  versus the actual which is you have no statistical proof of any kind of pattern whatsoever.

Much like the DNA evidence that is used to exonerate wrongly convicted felons, Trade Ideas servers watch every trade of every stock and baseline activity to identify if any pattern exists.  So here is the high level truth about patterns.  Most of the things you see when you look at a chart have no value at all.  Patterns exist as windows of opportunity that open and close.  You have to know when one is present to take advantage of it.

Trade Ideas users don't have to read the tea leaves and guess patterns.

Trade Ideas technology is by far the best piece of technology to help you identify when this window of opportunity is present. This post is the first in a series from Trade Ideas highlighting the importance of leveraging technology in the right way to truly capture alpha and not getting misled or distracted by the noise of a chart or your gut.

To download Trade Ideas and subscribe, or try the Free Trial, visit the Trade-Ideas product application download page.

Follow us on Twitter:

Like us on Facebook: TradeIdeasPro


Jeff White said...

David & Dan, I have used Trade Ideas since 2004 as best I can remember, so you know I like it. This post states support/resistance are not important, that they are just something we hope to derive value from and yet @tradeideasquant tweets regularly about broken support/resistance levels. Why?

I'm also curious as a pattern trader why the program contains so many pattern recognition features if your data shows patterns to be irrelevant?

Not every pattern plays out favorably, and neither does every data-driven trade. It's all about probabilities, so how is one method better than another? Seems to me they are just different approaches which can each succeed or fail based on the trader's discipline and abity to adapt.

You are both class guys and I look forward to seeing what you are about to uncover.


Jeff White said...

Maybe you are just saying the data confirms or negates the percieved pattern, in which case that is cool. I probably just commented too soon :)


TI Maven said...


The whole blog post was about your second comment. The data confirms the pattern. But also we are saying that recognizing patterns with your eyes is much harder than people think. The skills that you acquired over 15 years came through a kind of experience that is difficult to scale, which is why it is good that you teach. When we started Trade Ideas in 2003 we did not have the data to know what support and resistance actually looked like. We knew people needed to find them and drew on our own understanding. Our algorithms used our own visual recognition of support and resistance. But as we got to see the data underneath we started to see that many things people term as support and resistance simply don't have the mathematical value of actually being those points across a large data set. We are saying that now our technology, like the DNA in eye witness testimony allows you to mathematically test any technical level we track to see if there is a pattern or if it just looks like a pattern. The difference is one of success versus failure which is why we want to demonstrate that in our upcoming webinar on May 16th. We are redoing the way we present data and this is a big first phase :-)

TI Scott said...

Hi Jeff,

If you're not already registered for the webinar on Wed 5/16, please do. Here is the link to register:

Dan Mirkin and Dave Mabe are scheduled to demo the new version of The OddsMaker and discuss strategy optimization.