A design that included both microgenetic and cross-sectional components was used to examine 135 Slovenian children's acquisition of matrix completion proficiency and compare microgenetic and age-related changes on the task. The microgenetic analyses indicated that children's errors became increasingly variable shortly before they discovered the correct strategy, that the correct strategy became dominant quite quickly following its initial use, that improvements in matrix completion performance generalized to conservation, and that amount of learning correlated positively with IQ. The microgenetic/cross-sectional comparison, which involved contrasting the changes that occurred over seven experimental sessions with the changes that occurred between ages 6 and 7 years, indicated that the two patterns of change were highly similar.