Judging objectivity in the traditional game of Gobak Sodor remains constrained by its reliance on visual observation. This conventional system is prone to subjectivity and human error. This research aims to design and evaluate a player re-identification system. This study is specifically positioned as foundational research. It aims to provide a technical basis for developing future objective judging systems. A state-of-the-art approach combining You Only Look Once version 8 (YOLOv8) for multi-object detection and ResNet50 for feature extraction was applied in this domain. System testing demonstrated perfect performance. The model achieved 100% accuracy for Cumulative Match Characteristic (CMC) Rank-1 and Rank-5. Furthermore, the mean Average Precision (mAP) score reached 1.00. These results confirm that the proposed method combination is highly suitable for the traditional game domain. The system proved capable of performing deep feature extraction for each player. It was not limited to simple attributes like costume color. This research successfully provides a solid technical framework for modernizing judging systems in similar traditional games.