This study investigates the effects of overseas training on errors and L1 interference in English essays written by Korean university students. The subjects included 124 students who did not have overseas training and 43 students who had experiences of studying in English-speaking countries more than one year. They were asked to write a one-page self-introductory essay at the beginning of each semester. Errors were analyzed according to a list of grammatical categories. The essay length and error rates were compared. The study also makes prognostic suggestions for teaching writing in Korea by diagnosing the sources of errors. It was found that overseas training had positive effects on the essay length and error rates. It was also effective in reducing awkward expressions. The error rates directly attributable to L1 were 8.6% for the overseas experience group and 16.4% for the non-overseas experience group. Error rates attributable to L1 based on the Fluctuation Hypothesis were 45.7% and 51.5% for the overseas and non-overseas experience group, respectively. Students made persistent errors in omitting articles, prepositions, the plural -s, and in using awkward vocabulary and expressions resulting from translation of Korean. The data shows that special prognostic training is necessary to reduce the typical errors.