Really well done. I use this kind of graph with teachers and their students' assessment profiles. I had read the original post and am embarrassed to admit that my original thought was that tables would be just as sufficient as graphs in this case. I'm always happy to be wrong and it happens a lot.

It's interesting to see the bizarre shaped of 2007, as is 2005. Obviously big circles are ideal, small circles not so much.

I wonder what ND's historical data and also what top 10 teams' charts would look like (not a request for additional work).

I'm going to guess that 1 or 2 major indentations could be tolerated (producing the pac-man shape), but more than 1 major indentation on one side of the ball I would find it unlikely to be successful.

Savage, excellent job!

Another thought - if you anchored the maximum, you could overlay 'talent' rather than the second year on the graph by multiplying star rating by a common factor. For example, if you set your maximum to 25, you could multiply star-rating by position by 5 (which would still be fairly easily interpretable) and you could see the 'talent' by 'experience' interaction. I suppose the maximum possible would be closer to 40 than 25 (a 3-year starter heading into his 4th season on a team playing 13 games per year).