Review scores: Pointless?

There's been a lot of talk in recent months about game reviews. How they've changed. How they need to change. How good scores are "bought" with exclusivity rights.
Now comes this piece, about boiling down reviews to a simple percentage number, and how that's a bad thing. The writer makes a very good point.

The writer assumes that anyone visiting a site with game reviews with no actual rating would seem silly. Do you feel that way? Surely the prose can indicate how good the game is? Quite often you can read a review, and then be baffled by the score at the end because all the reviewer did was complain about flaws in the game, and yet still gave it 90%.

In short, do we put too much weight on those numbers at the end of reviews?