The study presents results of a tablet-based math game intervention to enhance early numeracy skills of children in Tanzania. Standard level 1 children (n = 122), attending a rural primary school, were randomly allocated to either intervention or control group. The intervention group participated in a daily intervention session for 46 days. Children's performances in number identification, quantity discrimination, addition, subtraction, and missing number tasks were measured before and after the intervention with randomly selected children from both groups (treatment = 30, control = 31). Score gains in the intervention group were substantially greater than those in the control group. In particular, statistically significant effects of the intervention were identified in quantity discrimination, addition, and subtraction tasks. Item-level analyses using Item Response Theory showed that addition and subtraction items involving regrouping and most missing number items were too difficult even after the intervention. The study also identified which games were played the most or least during the sessions from play-log data and analyzed associations between children's test performances and gameplays.