The purpose of the present pilot study was to evaluate the effects of particular educational math games in kindergarten on numerical skills in Grade 1. Our participants were 49 children with early math difficulties, 65 average achievers, and 48 high-achieving 5-year-old children. All of them played various adaptive math games for 5 weeks. To avoid the Hawthorne effect, 37 children engaged in non-math-related games whereas the other children played counting, comparison, or numerical working-memory games. The results indicated that the effect of working memory and combined counting and comparison games continued until January of Grade 1. At the end of Grade 1, about 87% of the children with learning problems that had participated in the working memory game in kindergarten performed average on arithmetic skills. The present pilot study shows that children at risk of math disabilities can benefit greatly from the use of certain games to prevent learning difficulties.