Although I stopped at that guy with the lego cube/icecube/and cardboard I mean how difficult is to understand sparse points in 3d space. Enough with the props. Just replace each player in the lineup with his nearest-neighbour.
Anyway. Was expecting a little bit more detail from the first guy but I guess he kept it simple in order to make it understandable for a wider audience.
Still, I think (just from pure prediction standpoint) the future is not in manual constructed features. Just like computer vision and NLP, the future is automatically extracted features. But then of course we are going into low-interpretability black boxes.
I would LOVE to get my hands on the real time tracking data that they use though
The last guy I skipped because I figured he was more about the the communication part (I am interested in prediction)
I mean how difficult is to understand sparse points in 3d space. Enough with the props. Just replace each player in the lineup with his nearest-neighbour. 