It’s looney to celebrate New Years, New Millenia, etc.  Every day counts equally in the march of time.  Just by arbitrary historical accident one of those days is called the first day of the year.

But it occurred to me when I wrote the post about the rise in stock prices last month that there is social value from coordinating our focus on arbitrary milestone days.  If someone presents statistics to you about the behavior of some variable over the course of a year, which would be more meaningful?

  1. Stocks rose x% from July 9 2010 to July 9 2011.
  2. Stocks rose x% from Jan 1 2010 to Jan 1 2011.

Subjectively, it is more likely that the dates for the first range were cherrypicked by the statistician to generate the conclusion.  Restricting attention to dates that have significance “outside the model” makes the exhibit more credible.