Can we, as a group, just agree that we're all sufficiently aware of the fact that yes, average (by which we all mean "mean", don't be pedantic) does not require that 50% of the referent group is above and below that number for all sets of data, but, since we're almost always talking about a more-or-less normally distributed data set, it's close enough that the deviation isn't really worth talking about?
Because I'm somewhat tired of seeing this exact comment every time the word "average" is mentioned.
Re:Ten predictions (2004)
by DoubleMalt on Friday February 15, 2013 @08:47AM
(#5226090)
If we take your very narrow definition, the chances of any programmer to be (exactly) average are zero, if we relax a little bit we're back at tautology.
Than take {1,1,2,3,3}. By any reasonable definition the average is 2, and the most frequent values are 1 and 3.
Re:Ten predictions (2004)
by fauigerzigerk on Friday February 15, 2013 @09:14AM
(#5226210)
No, it's not a tautology at all. For instance, one sensible interpretation of the prediction would be that > 50% of all developers are within a range of += 10% of the average performance score. This may or may not be true.
It's not my fault the definition of average is very narrow.
The definition of average[1] is not narrow. It could be mean, median, or mode or something else.
Consider this set of programmer "skill levels" where most (80%!) are above average: {10, 10, 10, 10, 1}