Strategy

Bear in mind that you should have enough standard stars, but not too many. It is better to have 5 observations each of 30 standard/extinction stars, than to have 3 observations each of 50 stars; in the latter case, we are spreading the same observational weight over nearly twice as many parameters. Because one observation is ``used up'' in determining the stellar values, only of n observations per star are available for measuring extinction. Thus a star observed 5 times contributes twice as much to the extinction determination as a star observed only 3 times, which in turn contributes twice as much as a star observed only twice. The planning program will help you choose the right number of stars.

Furthermore, the running time of the reduction program is roughly proportional to the number of observations, but to the cube of the number of parameters to be adjusted. Thus, the program will take almost 5 times as long to run if there are 50 stars to be adjusted than if there are only 30. The storage required is proportional to the square of the number of parameters, and this may limit the number that can be handled on machines with small memories. At some point, any machine becomes overloaded; there is a practical limit to the number of stars that can be adjusted in a single solution. The reduction program automatically excludes non-standard stars that have been observed only once from the extinction solution.

In sum, though one can hardly have too many observations of standard stars, one should not go beyond the optimum number of stars recommended by the planning program. Some observers who follow Hardie's [10] method observe a large number of standard stars just once each; this is extremely inefficient use of telescope time.

Although these questions ought to be considered before one goes to the telescope, some observers only consider them when trying to reduce data that have been badly distributed. Even a sophisticated reduction program cannot generate information that was not gathered at the telescope.