Recognizing hand gestures is a prerequisite for human-machine interaction, e.g., in gaming and surgery training applications. Compared to vision-based recognition, methods based on data gloves, e.g., using IMU sensors, have several advantages in flexibility and responsiveness. The goal is to quickly and accurately recognize hand gestures while minimizing the number of employed sensors. There are, however, several challenges, both in the combination of hardware and software solutions and in gesture design and respective recognition algorithms. In this study, we explore the advantages of using data gloves in gaming with a minimum number of sensors per finger. We demonstrate that users wearing the data gloves can accurately control an actor in a video game in perceived real-time.