Um, that video is clearly just a mo-cap test. Apparently Natal uses machine learning to interpret gestures, so it requires lots of data from lots of different people, moving in lots of different ways. Also good to see that they are tuning it for young kids. Just because it will work for...