The rest of this paper is divided in two Sections and a Conclusion.
In Section "Min-max and newsboy model" the reasons
to study the intermeans parameter are given. The first subsection
describes the newsboy problem and introduce our notations. A new formula
for the cost of uncertainties will be given that introduces the parameter
. The second subsection recalls the [8]
results concerning the family 
. The last
subsection shows that considering the family 
leads to very different conclusions.
In Section "intermeans parameter properties" , this
parameter is studied per se. The first subsection shows that
is another measure of the dispersion. The second subsection
compares with the usual measure, namely the standard deviation
. The last subsection deals with the sampling distribution
of . This is a key result since it must be shown that
can be extracted from historical data at least as well as the usual
.
The paper ends by a concluding Section, that summarize our results, followed by a short bibliography.