The present study evaluated the effects of a combined form of written corrective feedback (WCF) on English as a foreign language (EFL) students' writing accuracy. The combined WCF consisted of unfocused error-code WCF and focused metalinguistic explanation. Different forms of WCF were administered to two groups of Japanese EFL students in two consecutive years, and the effects of the feedback were compared based on the number of grammatical errors that the students made before and after receiving feedback. The original version (single combined WCF) provided metalinguistic explanation only once for each of eight target grammatical forms, whereas the intensive version (repeated combined WCF) provided metalinguistic explanation repeatedly. The results showed that combined WCF facilitated the students' accurate use of the target forms overall, and repeated combined WCF was more effective than single combined WCF, but its efficacy weakened over time. Repeated combined WCF had a positive effect on students' accurate use of verb tense and the avoidance of informal usage; single combined WCF had a significant effect only on verb tense. Repeated combined WCF also served to reduce the total number of errors, including errors for which no metalinguistic explanation was given, implying that coded WCF had its own contribution to the students' writing accuracy.