Comment: Statistical analysis of cycles.

(See in situ)


Statistical analysis of cycles.

Back in the 1980's I read, "The Profit Magic of Stock Transaction Timing" by J M Hurst. His idea was that the price of a stock today is determined 75% by trend, 22% by cycles, and 2% by randoms. I also picked the minds of people at the Foundation for the Study of Cycles while they still operated in PA.

Determining trend can easily be done by using some formula for a line and deriving the values for that formula by analyzing the data. For interest rates I use a simple straight line formula, but for data denominated in dollars I use a 3rd degree polynomial because that fits compound growth curves better. I use at least a data series of weekly data going back 50 years.

Cycles can be found using various statistical techniques after detrending the data. There are hundreds of cycles oscillating simultaneously and it is just a matter of identifying and testing each one to determine the probably that it is not random, and retesting as each new data point is added to the data series.

Randoms because they are random are small and have a tendency to cancel one another, so I ignore them.

The trend and cycles can be projected mathematically into the future once they are identified and measured. It never works out as projected because there is averaging involved in measuring each oscillation, and the next oscillation won't necessarily be average. Like a heart, they can occasionally even skip a beat. But for a buy and hold strategy it is sufficient to produce profits.

This technique worked well enough to get me into silver in late 2001, early 2002, and get me out just one month too early in 2011. But my silver projection showed a much quicker drop than actually has developed so far, but the direction has been downward since April, 2011 and my current projections still show we will go under $5 in 2014 and after a long trough won't be back above $20 until 2021.

For interest, up until 2007 my projections were closer than on any other market I follow, but the huge FED intervention in 2008 drove the rates down further than my method predicted and kept them from going up as the historic cycles would have done. So now my calculations show that we will go much higher in 2014 than we would have, had the FED not intervened. Had the FED not intervened my 2014 target for the 10 year treasury would have been 7.5%. When you push the pendulum artificially too far one way, it upsets the system and it must swing back too far the other to balance.

As for the stock market, I have been predicting a major move up since this time last year, but it has been slow to develop. I still show it exploding upward over the next 12 months.

My strategy over the last year has been to be short silver, short bonds, and long the stock market, which is what I am staying with. I have made money on the short silver position and the long stock market position, but have gone nowhere with the short bond positions.

I have overall done very well with this technique, but I have had a couple of > $1 million losses too, so it is not for the faint of heart. To me it is a great game to play, much more addicting than gambling which I don't do, and drugs which I don't do either.

"Bend over and grab your ankles" should be etched in stone at the entrance to every government building and every government office.