The light-curves of many variables, prepared from a large number of individual estimates by many different observers, have a lot of scatter, ranges greater than one magnitude being common, although it is generally accepted that the eye can detect differences of roughly 0.1 magnitude, at least for stars of moderate brightness. The question arises whether these gross discrepancies can be reduced. We think that this is possible if observers make better use of their observational capabilities.

For certain types of variable stars, e.g. long-period ones, the lack of accuracy may often be balanced by the vast number of estimates available. But for other stars, especially those exhibiting small-amplitude fluctuations, this process of combining observations leads to broad light curves that completely hide any possible fluctuations. Variable-star observers are familiar with these problems, but it is desirable to try to improve the accuracy that can be obtained.