I'm not thinking of predicting the next shot as much as estimating the center of the distribution being sampled. So for a group I need the means in x and y directions. From the mean point I can them measure the radius to each individual shot. Once I have that number for each shot I can calculate mean radius, or root mean square radius.