From: Harvey Newstrom (mail@HarveyNewstrom.com)
Date: Thu Mar 21 2002 - 06:12:24 MST
> steve wrote:
> >
> > http://www.guardian.co.uk/Archive/Article/0,4273,4373847,00.html
This experiment is so bad on so many levels, I don't know where to begin.
First of all, one girl, one astrology, and one stock analyst does not make a
statistical sampling. There is no way to know if this result is repeatable.
Secondly, having a little girl catch random pieces of paper is not a
scientifically recognized method for random number generation. This
obviously is a publicity stunt and not a real experiment to measure random
chance.
Thirdly, they claim that the girl catches the papers in a random manner. We
have no information about how the stocks to drop to the girl were chosen.
It seems that whoever chose these stocks to drop was making the good
choices.
Fourthly, all the girl's catches were winners. She caught no losers, and
there was no mention if the stocks not caught by the girl included any
losers. It may be that whatever source they used for stocks to drop to the
girl was the best predictor of good stocks.
Fifthly, the first description of the participants described a stock
analyst, but a later description describes "calculated, computer-backed
choice" instead. Who programmed the computer and what assumptions did it
make? Computers looking at previous results are notoriously bad at
adjusting for market changes. The article mentions that the girl avoided
the telecomm sector. If the humans and computers picked telecomm stocks
because they were the best performing, then random chance would be better
since it wouldn't favor the best stock which later failed. All this
experiment shows is that telecomm went down, those that chose telecomm
stocks lost, while those that didn't choose telecomm stocks gained.
I can't imagine that anyone could set up such a poor experiment, that anyone
would use a little girl as a random number generator, that the little girl
acting randomly never chooses a poor stock, or that this whole thing isn't
just a publicity stunt. I can't see how any real scientist could work with
an "experiment" like this. It is so invalid in so many ways, that I would
fail any high school student who tried to do an experiment like this.
-- Harvey Newstrom, CISSP <www.HarveyNewstrom.com> Principal Security Consultant <www.Newstaff.com>
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:13:03 MST