Feb

7

The stock market has emerged in last 5 days and it isn't captured, I dont think, by the normal things. Here are the high and low closes for recent days:

date  open      high      low        close
206   1454.0   1454.8  1447.8   1453.3
205   1451.5   1454.0  1448.1   1453.7
202   1452.2   1454.0  1448.1   1453.7
201   1446.3   1451.7  1444.0   1450.8

The average absolute change in highs, lows, and closed, from day to day, is 1. This has to be an all time non-holiday low. IBM also is trading at exactly 100 after swinging back and forth 4 times in last month above and below. What does it portend?

Not having any keys, although I do have the book by Ken Follett, I would like to consider some childlike questions about it. Others might think about this and the generalizations of same, I think, with value.

Sushil Kedia writes:

The lull before the storm. A single day's behaviour such as this would be dismissed as indecision.

Similarly, a second day would, at best, be termed market failed to get out of its indecision.

A third day again like this would make one tilt towards thinking the simplest of possible explanations, that a lull often is seen before the storm.

As distinct from any breakouts, which might not exist for a profit seeker, a simpler visualization appeals here. In ball games, from a football to a cricket ball, the point in the trajectory where a noticeable spin seems to be developing is a similar moment of quickly vanishing lull. Not a point of reversal, not a point of inflexion, just a point where the mistress will try to shuffle out the maximum number of players diving in either direction.

I clearly have no clue how I could translate this string of thought into a testable hypothesis. 

Kim Zussman writes: 

SPY, daily partitioned into 10d non-overlapping periods (from today's close) back to 2000; every 10d period checked standard deviation of daily closes, return for this 10 days, and return for next 10d.

Ten day return regressed against concurrent 10d standard deviation was negatively correlated (T=-1.9). Regression of next 10d return against prior 10d standard deviation and return showed positive correlation with prior standard deviation (t=1.8), and slight positive with prior 10d return (t=0.9).

Going to 5d non-overlapping, the current 5d standard deviation is 6th lowest of 352. The same regressions showed different results, with slight/NS correlation between 5d returns and concurrent standard deviation. The multiple regressions for next 5d return showed significant negative correlation with prior 5d return (t=-3), and slight positive/NS with prior 5d standard deviation (t=0.24).

So it looks like over short intervals, SPY returns related more to prior returns than volatility; but in longer intervals prior standard deviation is more important. 

Vincent Andres writes:

"The average absolute change in highs, lows, and closed, from day to day, is 1. This has to be an all time non-holiday low. IBM also is trading at exactly 100 after swinging back and forth 4 times in last month above and below. What does it portend?
Not having any keys, although I do have the book by Ken Follett, I would like to consider some childlike questions about it. Others might think about this and the generalizations of same, I think, with value."

Some related thoughts:


1.) K. Lorenz often put emphasis on the pair: stimulus and duration (and duration is often considered on spec list). Not surprisingly, in general, the longer the duration, the smaller the needed stimulus to provoke an identical reaction. Maybe wrong, but I wouldn't be surprised if duration were often a good candidate to explain our stats, residues, and even more.

On more elaborated stimuli, K. Lorenz and Tinbergen speaks about "triggering schema" (schéma déclencheur). This concept may be an appropriate frame for some of our stimuli. Tinbergen got the Nobel Prize with Lorenz and Frisch. R. Dawkins was a student of Tinbergen.

2.) Remembering the previous "trend" thread, we may consider non-trending phases as quite rare events. So, if the stimulus part alone is rare, this seems a condition propitious for the whole pattern being non-random.

Apologies, no counts (… not yet).

PS: "I have found the missing link between the higher ape and civilized man: It is we." K. Lorenz. 


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search