This book explores the complex interplay between test-takers' strategic competence, task complexity, and performance, focusing on Chinese learners of English as a foreign language (EFL). It introduces and empirically validates the Strategic Competence Model of L2 Speaking (SCMLS), the first framework in second language research to clearly conceptualise strategic competence in speaking, designed primarily for Chinese EFL learners. It also provides a feasible solution to the long-standing challenge of designing and assessing task complexity and multimodality in L2 testing by using Robinson's (2015) Triadic Componential Framework. Methodologically, the employment of quantitative hierarchical linear modelling and qualitative, ChatGPT-assisted coding to investigate these relationships offers rigorous and efficient approaches to data analysis for future studies. The book provides pedagogical implications for metacognitive scaffolding and task development for EFL speaking instruction. This book serves as a useful reference for language assessment researchers, applied linguists, teacher educators, test developers, and graduate students interested in second language speaking performance and strategy use in assessment contexts.