why is it better to look at the SAME fields each night?

[Stupendous Man <richmond@a188-l009.rit.edu>]

  Last week, I said:

>   If you can arrange it so that the 2 or 3 fields you follow each
>   night are _roughly_ the same (that is, there's an overlap of at
>   least one degree in RA from night to night to night), then your
>   dataset will be much stronger than if the "follow" fields are
>   different on each night.

  Tom asked me to clarify.  It's simple: suppose you have 20 clear
nights over the summer.  If you look at the same fields each night,
then you have very good sampling of the stars over timescales
ranging from a few minutes (from one image to the next on a single
night) to two hours (start to end of run on one night); then a gap,
and then good sampling again on timescales of 1 day to, say, 60 days.
If one is trying to find variable stars, this frequent sampling helps
for three reasons:

      a) it gives one more data on each star, so that one can 
              determine the instrumental scatter more precisely
              and thus notice variability with small amplitudes

      b) it allows one to determine the period of variability
              more precisely (for periods of a few to 60 days)

      c) it gives one a decent shot to see rare events, such as
              the eclipses of binary stars with wide orbits
      
  I think that it's better to have 20 nights of data on 3 fields
than to have 3 nights of data on 20 fields.  

                                              Michael