'The big picture' is that Kovalchuk did not improve the team at all 5 on 5 and perhaps hurt it. A dynamo shorthanded and a help 5 on 4 (although who knows how much of it he gave back through short-handed goals). Shootouts were a plus though again he was terrible at them before he came to NJ. It's hard to imagine him as a 2 wins/year player.
Tri - I value your analysis a lot, almost always I think that your evidence-based approach is correct. In this case, I'm wondering if the data you are looking at has been adjusted for quality of competition. I'd imagine Kovy draws all the hard matchups 5v5, allowing for some other players to play against 2nd or 3rd pair defenders. Do you know if the data has an opinion on that one way or another?