From: nospam@de-ster.demon.nl   
      
   Thomas Koenig  wrote:   
      
   > J. J. Lodder  schrieb:   
   >   
   > > Naval chronometers could have sudden jumps   
   > > in their going rates for various reasons.   
   > > Take one, and you hope for the best.   
   > > Take two and you go crazy when they start to diverge.   
   > > (which one is right, so where am I ????)   
   > > Take three and you can hope to identify the faulty one.   
   >   
   > "To measure once leads to knowledge.  To measure twice leads to   
   > doubt. To measure three times leads to statistics."   
   >   
   > (An unattributed Internet quote which seems to be a generalization   
   > of what you wrote).   
      
   Yes, but those captains had more than statistics to worry about.   
   Their main worry was a chronometer changing its rate   
   in an unpredictable way. [1]   
      
   Perhaps the background for this story was Royal Navy policy.   
   They would supply one chronometer to each RN ship,   
   unless the captain also owned one privately.   
   In that case they would supply two, on the ground   
   that a second one by itself would not give additional accuracy,   
      
   Jan   
      
   PS There is a lot of extra information in   
      
      
   [1] This is not uncommon in physics too. If you have two runs of   
   measurements, one with only statistical errors,   
   and one with also a systematic error,   
   then averaging the two series produces a result   
   that is worse than taking just the good measurements.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)