Just hit me that it'd be quite interesting given the advances in ML to see how well a computervision based system could be able to evaluate jumps,etc from these

Just hit me that it'd be quite interesting given the advances in ML to see how well a computervision based system could be able to evaluate jumps,etc from these sports. Once skeleton-detection is in, then a ML system given historical scores matching relative bone positions(with some adjustments for era proficiency, athlete physiology,etc) should be able to be fairly accurate in giving unbiased scores and be able to flag abberations in human judging.