Time Frame and
Randomness (H)
Too many investors focus on the short term as the media focuses their attention there. It sells more newspapers! How wonderful this world of short term randomness is to the media publisher!
You can be endlessly fascinated by these fluctuations and see any pattern that you may desire (and also create any explanation!) This is why day traders and the like never survive for very long. Of course a few get lucky and make a wonderful story for the financial media! You've got to sell those papers every single day.
The more information you give someone,
the more hypotheses they will formulate along the way, and the worse off they
will be. They see more random noise and mistake it for information. p144 The
Black Swan Also called the law of small numbers where
sample size is too small to make a conclusion. People make the mistake of overstating the
importance of small samples (in many cases one observation, “the market went up
today”.)
To clarify the matter, time frames are very important to H (randomness) and some behaviours are more predictable over the long term and others are not! By referring to p.194 in "Misbehavior of Markets" by Benoit Mandelbrot, you can see the different types of long dependence:
Focus on less ‘noisy’ charts and ignore those with ‘wild’ randomness.
When H > 0.7, long dependence is exhibited (easier to make a bet).
I identify this case as a chart fitting above the 45 wkma (week moving average). This applies in the case of oil, gold and CRB. Using the analogy of signal and noise, I consider this type of chart as a SIGNAL.
When H < 0.5, the chart is random and fluctuating.
I consider this chart NOISE.
To illustrate these concepts, I will return to my favorite chart of the moment, CRB (commodity resource bureau index of oil, corn, wheat, etc.):
CRB with 45 wkma
I think we all could agree that this chart has long dependence and is easy to place a bet on. The container I built is a 45 wkma that fits very nicely. So here H>0.7.
Now, let's focus on just 2006 to the beginning of September. Certainly, it was a very interesting year for commodities.
To compare like for like, I used a moving average that is the same fraction as above (the bull run has been going for 243 weeks, 45 wkma /243 weeks =0.185). For an 8 month period (2006), 0.185 * 34.666 weeks equates to a 6.4 week moving average.
It should be easy to make money right?
CRB 2006
Clearly, H < 0.5 no long term dependence and highly random. How could you honestly say that you could make money on this? Of course, those who bought in March and sold in July made out very well. Let's find these people and write a story about them. I bet I could sell a million copies! How smart these people are!
If you want to be a good investor and make money over the long run, it's important to eyeball a chart and ask yourself does this honestly have H > 0.7 ? Some charts are completely useless in the short and long run. Others like the CRB charts illustrate the importance of long dependence and making money over a period of time say 3-4 years.
The short term game is a fool’s gambit - at least with the CRB and I would say most markets. How do you tell? Look at the charts at different time frames.
One other point from Paul Slovic's
study of bookmakers: The increase in the information set did not lead to an
increase in their accuracy; their confidence in their choices, on the other
hand, went up markedly. Information proved to be toxic. p145 The Black
Swan