There are a bunch of things I think Carroll does really well, and I frequently find myself disagreeing with his detractors here on .NET, but I don't think this graph shows, well, really anything about how good Carroll is as a coach. The way I interpret that graph is that pretty much all NFL head coaches with a lot of games win about as many of those games as you'd expect them to win given the closing odds on the games their teams played. It looks pretty linear, and it's amazing how few coaches end up in the upper-left and lower right areas of the graph, where the coaches of teams that consistently overperform or underperform the betting odds would fall.
Even if a given coach is slightly above or below the best-fit line, I don't think that really says much about the coach one way or the other. Like
@Hawkamaniac, I think the main takeaway from the graph is that the betting markets do a pretty good job of predicting outcomes in the long run.