The current study examined whether preschoolers who are low achievers (LA) on mathematical tasks benefit morefrom a training programme that focuses on magnitude comparisons or ANS abilities (PLUS games) compared to games thattarget symbolic knowledge (DIGIT games).Twenty-four preschoolers played PLUS games and 21 children played DIGIT games 3 times per week for 5 weeks. Perfor-mance scores were compared to 25 typical control children who did not play any games. All children were assessed pre andpost-intervention on Test of Early Mathematics (TEMA), a computerized ANS task, the Give a Number task (Wynn, 1990) toassess cardinality and a counting and Digit Recognition task.The results showed that, although the DIGIT and PLUS groups performed lower than the Control group, both PLUS andDIGIT games improved mathematical abilities in LA children. These results suggest that there is a complex interaction betweenANS, symbolic, and formal mathematical abilities.