Video games are a multimodal cultural product, but localization research on them is still limited. This study takes Metro: Exodus as a case and examines Russian-to-Chinese localization with quantitative methods. Based on a self-built Russian-Chinese parallel corpus and grounded in Corpus-based Translation Studies and Translation Universals theory, it measures lexical richness, mean sentence length, part-of-speech distribution, sentiment, and dependency relations in the target text. The results show a clear tendency toward simplification in vocabulary and sentence structure, together with selective explicitation and emotional adjustment. Overall, the Chinese version keeps the text readable without weakening immersion.