well, you could certainly argue that sieges should be made more difficult so that 100% is harder to achieve, as I suggested more nationalist stacks to stop your advances and jump you little siege stacks would help.
I think I am promoting realism. Instead of having to completely carpet siege every inch of russia to get three provinces on your border, that should require like 15% war score or what ever scaling is determined, but make that 15% much harder to achieve. In this way, if you do manage to annihilate the odds and overcome all of the obstacles in your path, and totally siege your opponent, bringing their ruler to his knees, you can make proportional demands. I think this is fair and will improve gameplay and realism at the same time.
I'm with you all the way in saying that the war score cost equation ought to be changed, but a 1:1 relation is a bridge too far. The trick lies in finding out what exactly the scaling equation should be, in terms of math. I'm favorable to an X = Y, X2 = Y4, etc. scaling (where X is provinces taken and Y is the war score cost), but there's also X = Y, X2 = Y^2 as a possibility, plus different tunings of these.