This paper presents a simple yet effective approach for multi-object tracking (MOT) by leveraging an object detector without any specific training or optimization on tracking data. The proposed method, called Tracktor, uses the bounding box regression of an object detector to predict the position of an object in the next frame, effectively converting a detector into a tracker. Tracktor is extended with a re-identification (reID) module and camera motion compensation to achieve state-of-the-art performance on three MOT benchmarks. The analysis shows that Tracktor outperforms existing tracking methods in most scenarios, particularly in easy tracking cases, while dedicated tracking methods struggle with complex scenarios such as small and occluded objects or missing detections. The paper argues that Tracktor represents a new tracking paradigm that focuses on the core task of object detection and allows researchers to focus on the remaining complex tracking challenges. The method is online, robust to occlusions, and can be applied to various tracking scenarios without requiring tracking-specific training or optimization. The results demonstrate that Tracktor achieves superior tracking performance compared to existing methods and highlights the remaining challenges in multi-object tracking.This paper presents a simple yet effective approach for multi-object tracking (MOT) by leveraging an object detector without any specific training or optimization on tracking data. The proposed method, called Tracktor, uses the bounding box regression of an object detector to predict the position of an object in the next frame, effectively converting a detector into a tracker. Tracktor is extended with a re-identification (reID) module and camera motion compensation to achieve state-of-the-art performance on three MOT benchmarks. The analysis shows that Tracktor outperforms existing tracking methods in most scenarios, particularly in easy tracking cases, while dedicated tracking methods struggle with complex scenarios such as small and occluded objects or missing detections. The paper argues that Tracktor represents a new tracking paradigm that focuses on the core task of object detection and allows researchers to focus on the remaining complex tracking challenges. The method is online, robust to occlusions, and can be applied to various tracking scenarios without requiring tracking-specific training or optimization. The results demonstrate that Tracktor achieves superior tracking performance compared to existing methods and highlights the remaining challenges in multi-object tracking.