About OnCourt Insight

How Stroke Recognition Works

Our stroke recognition uses advanced machine learning to process acceleration, rotation, and other smartwatch sensor metrics—distinguishing stroke types with high accuracy.

• Data Collection: Accelerometer and gyroscope data capture each movement; the model is trained on thousands of strokes.
• Shot Classification: A custom neural network categorizes each movement as a forehand, backhand, or serve—including spin variations.
• Continuous Learning: Accuracy improves over time as more data is gathered.

Data Privacy and Security

All personal and match data is securely stored in encrypted databases on Google Cloud. We follow industry best practices for data protection, and your data is never shared without your consent.

We design for transparency and control, so you always have visibility into the data collected and how it’s used.