[extropy-chat] the structure of randomness

[The Avantguardian]

--- gts <gts_2000 at yahoo.com> wrote:
> I should qualify this a bit. The supposed
> non-monitonical ("toothy") shape  
> of single histograms is part of the alleged effect,
> but I find that much  
> less interesting

I agree with you that the non-monotonic nature of the
histograms is less interesting than their alleged
periodicity. In fact the expectation of monotonicity
in presumably random distributions is, I think, a
mistake similar to the fallacious law of averages. For
example if I start to flip a coin and tally the
results, it is true that as I flip the coin the ratio
of heads to flips will start approaching 1/2. The
fallacy however is the unfounded belief that this
number approaches 1/2 monotonically the way a
mathematical limit might approach 1/2 with each result
bringing me closer to 1/2. Instead this is not the
case at all. After 10 flips for example, the ratio
could very well be exactly 1/2. Then I might flip it
10 more times and get heads everytime. This would take
the ratio away from 1/2 to 3/4 and lead to
non-monotonicty. Thus the fallacy is that more trials
ALWAYS brings one closer to the expected average. In
fact if you think about the more trials you have the
more likely you will get larger devaitions from the
mean. For example it is far more likely that you will
have at some point had a run of 50 heads after 10,000
tosses than after 100.   

> than the claim that these patterns
> persist and recur  
> periodically at 1, 27 and 365 days.
> 
> However the more I think about it, the more
> convinced I am that Sarfatti  
> is right to think Walker refuted it.

I have read and thought about Walker's claim to refute
Schnoll's work. While his explanation for the increase
of similar histograms at +/- 0 hours is plausible
(i.e. sampling bias of counting of 2 hrs worth of
histograms as 1 hrs worth due to a systematic error) I
don't buy Walker's explanation of "routine computer
housekeeping" for the increase of similar histograms
that Walker himself sees in his own data at +/- 24
hours.

 First of all, if it was such a thing, you would think
that Walker could easily turn off the tape-backup,
virus scan, or whatever else he presumes is causing
his computer to oversample similar histograms in those
regular time intervals.

Second, this explanation could not possibly hold true
for the spikes seen at 27 days and 365 days as these
are not at precise intervals of 24 hours. ie. There
were decimals that I don't recall after these numbers
thus, the idea of some computer process that occurs
with a regularity of 24 hours could be discounted.

Third, Schnoll first noted and started observing this
phenomenon 40 years ago, before Unix and any bias
introduced by periodic Unix processes could have been
introduced into the data.

Therefore, I do not believe Walker has refuted
Scnoll's data, although he might have pointed out a
source of error for the immediate time measurements,
he does in no way refute the periodicity of Schnoll's
data which is the truly interesting part of it anyway.

I understand that there is an urge to ignore anything
that does not fit into ones tidy little paradigm, but
all true progress of the paradigm depends on analyzing
the anomalies and not ignoring them.
  

The Avantguardian 
is 
Stuart LaForge
alt email: stuart"AT"ucla.edu

"The most beautiful thing we can experience is the mysterious. It is the source of all true art and science. He to whom this emotion is a stranger, who can no longer wonder and stand rapt in awe, is as good as dead: his eyes are closed. . ."

- Albert Einstein, "What I Believe" (1930)


	
		
__________________________________ 
Yahoo! for Good - Make a difference this year. 
http://brand.yahoo.com/cybergivingweek2005/