BoulderGrade
ML-driven grading for training boards
(currently Kilterboard only)
How BoulderGrade Works
1. Upload Your Route
Open the Kilterboard app and take a screenshot of the route you want to grade. Either crop it or leaf the screenshot untouched, as the system will find the route in the image. Upload the image using the drag-and-drop area or the file browser. If the system does not recognize all holds or takes too long to compute, this might be dues to early development bugs or too many system requests. Feel free to cantact me as I am always happy to help and improve the system.
2. Specify the Wall Angle
Enter the angle of the Kilterboard wall (currently only 40 degrees). This is important for accurate grading, as steeper walls generally result in harder routes. The standard Kilterboard angles range from 0° (vertical) to 70° (very steep).
3. AI Analysis
My machine learning algorithm analyzes your route by identifying the holds, their positions, and the overall pattern. It compares your route with thousands of other graded routes to determine the difficulty.
4. Get Your Grade
Within seconds, you'll receive an estimated grade for your route along with a confidence score. As it is still in development, the system might not always be accurate. Furthermore, accuracy is a quite subjective thing in climbing, so the grade might not always be what you expect.