This paper explores the hypothesis that livestock market reporters tend to err towards the middle of the actual range when estimating carcass weight and fat depth. Trial data for pigs support the hypothesis and indicate that reporters' estimates may be significantly biased in some extreme weight and fat classes even when the average errors taken across all classes are trivial. Significant distortions in the reported premiums or discounts between some fat and weight classes are implied by the errors in estimating fat and weight.