« Tropical Views | Main | Tropical Views »

November 28, 2006

Understanding Signal and Noise

Wikipedia tells us the signal-to-noise ratio is an engineering term for the power ratio between a signal (meaningful information) and the background noise. More informally, it is used to describe useful information (signal) and false or irrelevant data (noise).

 

Michael Mauboussin, Chief Investment Strategist for Legg Mason and author of the book "More Than You Know: Finding Financial Wisdom in Unconventional Places" describes signal and noise this way:

 

Time horizon is a crucial consideration in any probabilistic field. In these systems, short-term results show mostly noise—the noise-to-signal ratio is very high. Over time, the signal reveals itself, and the noise-to-signal ratio drops. Short-term investors dwell mostly in the world of noise.

 

Let’s go back to our simple coin-tossing example to demonstrate this point. Exhibit 8’s left panel is the result of a 20-toss trial, and shows that 35 percent of the tosses came up heads. (We simulated these results with a random number generator). The panel on the right continues with the next 80 tosses in the series, and shows that the ratio settles very close to 50 percent over 100 flips. Even though we know the long-term signal is 50 percent, short-term noise can deviate substantially from long-term signal.

 

Many Snider Investment Method™ practitioners have come up with for ideas to modify the Snider Method. The challenge is that almost all of them are based on emotion, intuition or very tiny data sets. When we deal with small sets of data, as in the case of the first 20 coin flips, we are usually basing our conclusions on noise rather than signal.

 

Recently, a graduate said to me that he had bought three stocks that declined in price. All of them had a high P/E. He wanted to avoid high P/E stocks thinking this would reduce the likelihood of declining stocks in his portfolio. Of course, three stocks is not sufficient to test the validity of an idea.

 

In reality, there is no strong correlation, that we have been able to determine, between high P/E and falling stock prices. Excluding the higher P/E stocks would eliminate a number of our better performing stocks without making any meaningful improvement in risk. He was basing his idea on noise, not signal.

 

It is in everybody's best interest to make improvements to the Snider Method. That is beyond question. The only question is how can we marshal the resources required, which are considerable, to do it properly?

 

This is the idea behind making the Snider Method "open source" if you will. The key is going to be creating enough data to be able to test ideas across very large sample sets to remove the noise and get to signal. This is no small task and beyond the resources available within the company. Hence, the idea of making the method open source to its users and inviting them to participate, for the collective benefit of everyone, in advancing the method.

 

What I am hoping to accomplish at the Open Space meeting on February 3rd is to lay the foundation for working as a community toward improvements in the method. The purpose of the Open Space meeting is not to float the ideas themselves, but rather to create some organization among those who are interested.

 

I am hoping we can eventually use the collective time, wisdom and talents of the many smart people who use the Snider Method to gather enough data, organize it, and create a "test bed" where everyone helps build a sophisticated tool for empirically testing all modifications to the method. It would be up to the community to agree on what constitutes empirical proof and create a process for submitting improvements that have met that proof. Once we have a framework, people can pursue the area that interests (or bugs) them the most -- selection criteria, money management rules, rolling options, etc.

 

We have already had quite a few people sign up for the Open Space meeting even though it is still pretty far away. You can read more about in the original post. I also sent the post to everyone as an email announcement.

 

Call the office to register. The number is 214-965-9950 or 866-952-0100. They will take your credit card over the phone.

 

If you have any thoughts, comments or questions about signal to noise or the open source concept, please feel free to leave them below.

 

SOURCE:

 

Michael Mauboussin, "Long Term Investing is a Short Term World" Mauboussin on Strategy, 18 May 2006

http://www.leggmason.com/funds/knowledge/mauboussin/mauboussin.asp#

 

Kim Snider, Kim Snider Financial Communications, Chronim Investments and/or Snider Advisors make no representation that the information and opinions expressed are accurate, complete or current. The opinions expressed should not be construed as financial, legal, tax, or other advice and are provided for informational purposes only. Call 866-952-0100 to request the Snider Investment Method™ Owner's Manual, which includes a description of the Snider Investment Method, investment objectives, risks, suitability and other information. Please read and consider carefully before investing. All investments are subject to risk including possible loss of principal.

 

Posted at 09:55 AM in Events, The Snider Investment Method | Permalink

Technorati Tags: Kim Snider, Legg mason, Michael mauboussin, More Than You Know, open source, open space, signal-to-noise, Snider Investment Method, Snider method