I am skeptical like Wouter. The concept is to build some compliance into the rudder system to accomodate differences in the "best" or optimum angle of attack for each blade relative to the other blade. As Wouter pointed out the article uses a point solution with some oversimplistic assumptions to draw a broad conclusion about potential performance improvement. The most reliable way to validate something like this is boat on boat test results with skippers switched for half of the test runs.
By logical extension, if we made the platform so one hull can move relative to the other, less energy would be lost in going through waves (theoretically at least). But it obviously dosen't work that way in the real world. Another unanticipated effect could be introduction of rudder vibration if the spring constant/mass of the more flexible rudder system is close to the vortex shedding frequency of the blades.
Speaking of compliance in a design... the windsurfing guys have done a great thing by designing the upper third of the sail to be floppy. It blows off in a gust and reduces the angle of attack. A sail like an automatic transmisson in you car. I mention this to point out that lifting surface compliance can be very effective in specific applications.

Kevin