Re: Benford's Law

[Myron Tribus <mtribus@home.com>]

I would expect Benford's law to apply only to data which had been 
derived through the multiplication of different numbers representing 
concepts which have some meaning to the data.  Thus, if the 
proprietor of a hardware store simply sets prices at $0.99,  1.99, 
etc., this would not qualify.

Myron Tribus,  350 Britto Terrace,  Fremont, CA 94539
Ph:510 651 3641  Fax: 510 656 9875   e-mail: mtribus@home.com
There is no such thing as an immaculate perception. What you see 
depends upon what you thought before you looked.

At 10:33 AM -0500 12/3/00, Alla Linetsky wrote:
>The article in New Scientist clearly cautions against bounded, limited
>and otherwise non-naturally occurring data. Even in the hardware store
>example, where the fraud supposedly occurred, how many items are
>likely to be priced at 99 cents? How about $1.99? $9.99? Our
>retailers' penchant for manipulating our feelings (and wallets) by
>fixing prices would make Benford's law unsuitable to small retail
>operations.
>
>As for weather, especially if one uses the Fahrenheit measure, how
>often does the temperature in an average American city exceed 100F or
>drop below 20F? Not 30 percent of the time, I would guess, although
>the law may well hold for Celcius degrees.
>
>Alla Linetsky