This, by Rick Hertzberg a while back, is looking better by the day:

…Centrism sometimes makes sense as a tactic or a strategy — in other words, when it’s a synonym for compromise. But it has no merits as a tool for policy analysis. I suppose you could argue that good ideas occur on a sort of left-right bell curve and that, therefore, an idea is statistically more likely to be located at the top of the curve, i.e., in the middle. But evaluating the merits of an idea on that basis would be like evaluating the literary merits of a novel based on how close its number of pages is to the average for all works of fiction.

Dogmatic centrism not only puts you at the gravitational mercy of whichever side is prepared to move furthest toward its own extreme, it also obliges you to reject certain ideas automatically, without any analysis except spectrum analyses. That’s brainless, and the point holds whether or not you agree with Matt [Yglesias] on this particular issue…