Blake is nothing more than a role player, but IMO he is a good role player.
Remember how everyone was enamored with Shane Battier, calling him a glue-guy whose contribution is not really shown in "standard" statistics?
Maybe, just maybe, Blake is our glue guy?
Going a little bit off-topic and discussing +/- vs. Win% -
+/- is stat with a lot of noise. For example, if your team is consistently -3 worse with you on the floor for 10 games, but you run into the New Jersey Nets and blow them by 40 points with
you on the floor - you have a positive +/- - but it does not really reflect your contributions or how the team plays with you. Win% is less volatile to these noise conditions.
The criticism that both +/- and win% can have thrown at them - is that your stats might be skewed because of who you play with. For example, last year our best win% players were Roy (around 69% if memory serves), LMA (around 65%) and Blake (around 65%). The question you have to ask at this point - is Blake's win% so good because he plays so much with LMA and Roy - is a valid question.
But, this year, with Blake's Win% so much better than anyone else on the team - it is pretty clear that we are consistently a better team than our opponent when Blake is on the floor. Why? I don't know - but these are the numbers.
Just to illustrate how crazy his stat is - I did look around at some of the rosters of the better teams in the league - trying to find other players that have such a high win% - the only two I found (I ignored bad teams) were Al Horford at 91% and D-12 at 81%
Honestly - this is such an outlier as far as performance, especially when you consider how bad his individual stats are - that it is a very very interesting thing - but it is starting to make sense, to me, why Nate played him so much this year. I do not understand what it is that Blake is doing that makes us play better consistently when he is on the floor - but I can understand why he plays as much as he does based on the results.