Good lord man. Well done. In short, you factor analyzed a bunch of relevant tennis variables and found 4 types of players. Then you looked at how each type fairs against the other type. Extremely impressive work. Disappointed I did not see any type - type winning percentages that were convincingly greater than chance. Perhaps different clusters?