On 16 April 2002 I said :

  I have expressed concern recently over the generally
  poor quality of the pointing as described by the rms residuals
  in any data set -- see the minutes of any JCMT Operations meeting from
  the last couple of months.
  My feeling is that our advertised performance of 1.5" rms in each
  coordinate is not being achieved at the telescope, and that
  'recently' the figure has been more like 2.0" or more. I have therefore
  plotted below the rms residuals for all SCUBA datasets with more
  than 10 logged pointing data since 1997, expecting to find some
  deterioration in pointing quality in recent times :

I have displayed the plot here rather small so as to not lead your eye -- and you may click on it for a full view -- but maybe not yet !

The plot shows the rms of the azimuth pointing residuals in each dataset -vs- time (date) in the upper part of the plot, and the elevation rms -vs- time in the lower part. The azimuth data show, I think, a general reduction and tightening of residuals at about the half-way point in this diagram, but then abruptly exhibit a qualitative deterioration about 2/3 of the way through the plot. This is mid-late August 2000, identified on the plot by a red line. This coincides precisely with the introduction of the 'new' (at that time) TCS.

There is no corresponding change in the elevation residuals.

It was at first somewhat embarrassing to produce this plot almost 2 years after the event, and to see that my suspicions should have been raised much earlier, but it is also probably true that such an interval is necessary to generate sufficient data to allow this qualitative change to manifest itself - and maybe even then its reality will be considered somewhat in the eye of the beholder.

However, I do feel that suitable statistical processing -- eg monthly means -- would make the deterioration quantitatively convincing also.