A criterion for evaluating the practical advantage of spatial error control in single plant selection for lowly heritable traits, by either field preparation for environmental uniformity or statistical adjustment using spatial information from surrounding plants, was defined based on the principle of achieving the greatest success with the same long-term resource investment. The criterion was determined in the context of mass selection under practically possible values of the variables concerned, leading to the prediction that spatial error control will be rewarding only in some plants where the cost for error control is low compared to that for care management to raise the test plants. Error control will be of little use in other plants; improving the precision of selection when the population size is reduced is not rewarding. In cross-fertilizing cereal crops, error control for yield is predicted to be useful when the selection is performed in only a few cycles (generations), but the population size is important when the selection is performed in more cycles to best explore the genetic potential of the target population. In self-fertilizing cereal crops, spatial error control will in no case be cost-efficient. In both cross- and self-fertilizing crops, the population size becomes more important with a higher magnitude of the desired genetic advance.