Re: Which prices are fake?

From: Martin Penderis (tafeltennis_at_mwpenderis.mailshell.com)
Date: 09/23/04 

Date: 23 Sep 2004 10:53:27 -0700

Bob Wheeler <bwheeler@echip.com> wrote in message news:<414C8290.7090501@echip.com>...
> Of course Benford's law doesn't help here, but as a hint, it suggests
> that one should look at the digits, and the trailing digits suggest that
> the first series is made up -- too many 0's for randomness. I don't
> think I would say the same, had to told us that the first list was the
> result of human scale reading, since the trailing digits for such tend
> to mass on 0, 5, and 7.
The first list was taken from a newspaper of last week (16 September).
Your mention of 0, 5 and 7 is interesting. If you mean they are found
very often in the very last digit, then I believe that it is due to a
human tendency to round the prices to the nearest 5 cents. I cannot
explain the 7.

Benford's rule is useful with the first few digits and especially with
the very first digit. Using expected frequencies and given frequencies
for the first digit the chi-squared values are 8.743 for list 1 and
19.980 for list 2. The tail-end probabilities are 0.364 and 0.010
respectively. For a level of significance of 0.05 the critical value
is 15.507. The number of degrees of freedom is 8 for both lists. With
your observation in mind, the rule would not work on the last digit.
Whether one uses these results or this method to reject a list as
genuine or not, the fact is that the second list was made up.

Relevant Pages

  • Re: An uncountable countable set
    ... You intermingle numbers with digits. ... What is the tenth path in your tree? ... it holds for *all* lists. ... If all edges terminate, also all paths terminate, and both are ...
    (sci.math)
  • Re: Which prices are fake?
    ... The expected frequencies are too small for the chi squared ... There are a number of psychological papers that show humans tend round ... to the odd trailing digits 0,3,5,7. ...
    (sci.stat.math)
  • Re: abundance of irrationals!)
    ... contains only finitely many digits. ... That is only true for physical lists, ... If you mean that the decimal representation of 1/3 has MORE digits than ...
    (sci.math)
  • Re: How many digits is pi computable to?
    ... does not have anything to do with any of the lists ... Adequate natural language is available but you are ... Is pi computable to 0 digits? ... digits, for this entry. ...
    (sci.logic)
  • Re: Bash/grep -e filtering quickie question
    ... I indeed have 2 lists, and I wish to parse one field ... of one list with all the digits in one field of the other, ... > use sed to insert spaces, then allow the shell to split it up into ... > you want to do a single grep for all the digits at once. ...
    (comp.unix.shell)