Cognitive load theory (CLT) uses working memory resources depletion to explain the superiority of spaced learning, predicting that working memory resources will be less taxed if there are resting/spacing periods inserted between learning tasks, in comparison to learning from the same tasks in a single session. This article uses the working memory resources depletion effect, as a factor, to investigate the hypothesis that delayed testing would show superior results to immediate testing on math tasks for primary students in Singapore, as participants' working memory resources might be restored because of the resting between the immediate and delayed tests. Results confirmed higher performance on the delayed test than on the immediate test, as well as more working memory resources available for the delayed test.