Four observations of the same star field were obtained at different times (even across different observing runs). For each star, the rms spread in the reduced (MTPIPE) measurements was determined. The stars were then put into 1/2 magnitude bins and the average of those errors within a bin was plotted as a dot.
The upper line is the error assigned at that magnitude by MTPIPE. The lower (usually dashed) line is a similar prediction from J. Gunn.
