Here's a quick example to start.
2000 USO final
Safin - 32.3%
Sampras - 21.7%
2001 USO final
Hewitt - 27.8%
Sampras - 14.2%
Nothing too surprising here. Most people who have seen these matches would say, I think, that Safin's level of play was higher than Hewitt's. Not that Hewitt's play was low -- and these figures give a sense of how high it was -- but still, Safin's performance is regularly named as one of the best of the Open Era.
And when you watch the matches you can see plainly that Sampras runs out of gas in '01. Again these figures give some statistical idea of how far his quality dropped.
Nothing too controversial here, just an opening example of how to use AM's to study matches within a controlled context.