System and Discipline

"A Dash" is a blog about a book, written one page at a time.  It is an interesting process, since few people  in the intended audience for the book are even reading the blog.

Background

The book is aimed at the intelligent individual investor who is trying to use various methods and resources.  These are people much like our current investors (our friends --a very intelligent group), but who abandoned stocks for real estate after the 2000-01 experience.  Many of these people are still not actively invested.  We love the feedback from readers, via comments or email, but we do not expect to gain investors from this process.  Obviously, we hope that (ultimately) readers of the book will invest with us!

One concept that we have emphasized is determining a system that fits one's time frame and sticking to it.  This requires understanding and following a method.  The right method depends upon the time frame.

A few weeks ago we started to illustrate this by revealing our intermediate-term market positions as reported to the TickerSense Blogger Sentiment poll.  Many participants do not make their positions public, although a little calculation shows that most bloggers have been wrong.  CXO Advisory analyzed the data and found the poll to be a slight contrarian indicator, although not as much as other sentiment indicators.

When we started this process we knew two things:

1. Vince's TCA model was sound.  This was based upon our testing, designed to avoid the typical backfitting traps.
2. There would be periods where we would look foolish.  This was obvious from reviewing the simulation results, which showed some thrashing around and false signals.

These are typical of any model.  If your testing does not show it, you have probably made a mistake!

The most recent negative call looked bad for several days.  Many traders following the TCA model would have abandoned it, because they use short-term results to determine what is "hot" or working.  They would make the mistake of the blackjack player who notes the fall of the cards for a hand or two or believes in hot and cold dealers.

They would have quit at exactly the wrong time.  Our fund that tracks that model was actually up a bit today, and is beating the S&P nicely since we have been revealing the signals.  Our fund that blends long-term fundamentals with TCA was a loser today, but only by half of the market loss.  Our long-term positions in individual accounts traded with the market, since the discipline there keys on overall valuation and stock picking.  While that method was the biggest loser today, it has been the strongest performer over the last nine years.

Conclusion

The point of this article is not about advertising our approach, since our potential customers are not even reading "A Dash."  Instead, it is a message to every investor and trader that success comes from four important steps:

1. Discover a method in which you have confidence.  This should be based upon  sound testing.
2. Review your test data to see what bad streaks look like, and prepare to live through them.
3. Manage position size.  Many of the hedge funds that are blowing up have used leverage to magnify returns in most months, while ignoring what can happen in the worst months.  Their positions were based upon what they hoped to return rather than what they were willing to risk.
4. Stick to the discipline of your method.

Meanwhile, we wonder how investors using home-grown systems, as pushed by television advertising, have been doing during this turmoil....

Comments

[Larry Nusbaum]

"These are people much like our current investors (our friends --a very intelligent group), but who abandoned stocks for real estate after the 2000-01 experience. Many of these people are still not actively invested."

That's me. I am about 90% real estate. In the stock market, since 2000/2001, I have emphasized value over growth, small cap over large and much broader diversification. I have also added some shorts and some hedges while focussing on the precious metals mining stocks.

But, Jeff just because someone responds to the Tickersense poll one way doesn't mean that person invests the same way. For a long time I was bearish but remained fully invested.

[Josh Stern]

Hi Jeff, I'm one of those individual investors you ask about. The system that comes most naturally to me as an intuitive process is deep value investing with a nod to fundamental momentum and industry based relative strength (where I'm early enough to the party). I find that individual investors using fundamental analysis have a much greater chance of doing well if they stick to stocks that are not widely followed (the reverse is probably true for individuals using mainly TA). So I mainly end up in small cap value. I had a good week because that sector was up about 5%.

[oldprof]

Jimmie-

If you use the Google search window on the site you will see several topics on developing trading systems. The link to the TCA model development illustrates a good method. The post where we talked about Ralph Vince and his work does the same.

Most commonly, you have to avoid using all of the data or you have no out-of-sample test. You also cannot use the same out-of-sample too many times or it becomes part of the development data.

Good luck with your work. What a great opportunity!

Jeff

[oldprof]

David-

Thanks for our comments and the helpful links!

Jeff

[Jimmie]

You mention "backfitting traps." I am a student and have spent this summer doing some modeling for a local hedge fund. I am curious about what these traps may be. Can you elaborate?

[David Merkel]

Excellent post. Investors have to know what their edge is first, time horizon second, risk tolerance third, and the amount of work that they want to put in fourth. My best statement on this were the pieces that I wrote for RealMoney on using investment advice.

http://www.thestreet.com/p/_rms/comment/davidmerkel/10244269.html

I recently posted at my blog on my recent underperformance; it was a good exercise to go through. More investors need to analyze the negative side of their performance more. It hurts, but it makes us better.

http://alephblog.com/2007/08/07/dealing-with-underperformance/