An analysis is made of data obtained in a comparison of the color differences between faded and unfaded Blue lightfastness standards with Grey Scale steps, which showed that the ANLAB and CIELAB formulae result in color differences up to four times greater for the Blue standards than for visually equivalent Grey Scale steps. It is shown that this is due to an underweighting of the lightness differences against the chromaticity difference by a factor which depends on the relative amounts of each difference component. A formula is proposed which shows a near-perfect correlation with the experimental data and thus makes it possible to measure the degree of fading of any Blue standard in Grey Scale steps. When this formula is tested against Morley data, which are based on perceptibility testing, the correlation is worse than for the ANLAB or ClELAB formulae. The reasons for this apparent contradiction are discussed. -- AATA