The question of the effects of ability tracking remains unresolved even after decades of research. This is also due to the fact that specific regulations for the stringency of implementing differentiation according to ability and achievement have not been taken into account. The issue is the conditionalization of the effects of ability tracking within differentiating systems. The question is analysed using the example of the German federal states: There are clear differences in the regulation of differentiation (binding nature of recommendations and/or stronger organizational control of schools) and with the "National Educational Panel Study" (NEPS) a data set is available that contains all the information necessary to identify the effects. Three sequentially related aspects of the effects of stringency are analyzed: Sorting and educational attainment, structuring of school classes and educational achievement in secondary school. For educational attainment, there was an increase in achievement at the end of elementary school and greater equality of achievement at the transition without an increase in social inequality. The structuring of school classes showed a decoupling of cognitive and social segregation. And for achievement in secondary school, was an increase in the effects of cognitive composition without an increase in the effects of social background or social segregation, especially in the lower performance areas. This is practically the opposite of the assumptions known from most international comparative studies, according to which strict differentiation does not improve achievement, but only reinforces the effects of social background.