Understanding and identifying tactical patterns in football is important for performance analysis and strategic decision making. This work presents a modular method that combines time series feature (TSF) extractors with clustering and cross-analysis to identify and evaluate offensive and defensive patterns. Using tracking and event data from professional games of the Deutsche Fußball Liga (DFL), we extract permutation-invariant spatial features from sequences that precede key situations. These sequences are processed with catch22 and QUANT TSF and clustered via k-means. By cross-analyzing offensive and defensive clusters, our approach estimates the probabilities of success and predicts outcomes in goal-related situations. The evaluation shows that catch22 with selected spatial features achieves strong predictive performance and produces interpretable patterns. Therefore, our paper offers a framework for tactical analysis and could help coaches and analysts identify effective or vulnerable matchups.