Some empirical results:

http://www.moth-sailing.org/download/CSYSPaperFeb09.pdf

"The strut analysis brings home several points:

• Thinner is better. Certainly, in straight ahead tests thinner foils performed better, the limit being structural. It is unknowen if this trend would hold if yaw were considered.

The lifting foil wave drag coefficient is gratifyingly consistent, but is on the order of half the value predicted by Hoerner. Certainly the lightly loaded rudders show little wave drag penalty for operating near the surface, but the maneuvering implications for operating very near the surface are grave indeed. The more heavily loaded daggerboards gain about a pound of wave drag operating near the surface, but this is offset by the reduction in strut drag. This result confirms what Moth sailors have discovered; the higher you fly, the faster you go.

The induced drag is about a quarter of the daggerboard T-foil drag, but only about 5% on the lightly loaded rudder.

Summary Drag Breakdown:

Figure 23 summarized the drag breakdown for the six T foils tested. Small gains could probably be realized by pushing material limits harder and making thinner foils."

 

Findings as a percentage of total drag for each foil: Strut wave and spray drag is of the order of 3-4% across all foils. Dagger foil wave drag is roughly 4% and rudder wave drag, around 1%.  It is for good reason that I have focused upon thinner foils and not yet looked into wave drag theory.  Junction, section and strut section drag account for over 80% of total rudder drag and almost 70% of dagger foil.

The absence of empirical results for the bi-directional proa sections means that this work is highly speculative, but that's no reason to dismiss its potential.  As Luc has observed, apart from Kevin O'Neil's qualitative analysis, we don't have much to go on.  If there is any more comparative evidence, I'd like to peruse it.  If science had taken the dismissive approach throughout human history we wouldn't have got very far.  The scientific method entails attempting to falsify a thesis - creativity disciplined by scepticism.  Perhaps other sections will be explored.  I have some in mind.

The new HP rudder mechanism is premised upon bidirectional foils, so we may as speculate, theorise and empirically test these things.  I posted here in order to test my ideas, which on balance has been productive.

<<Also the L/D you are quoting is for 2D flow.  It will be much less when you start considering realistic spans able to support tonnes of boat.>>

Initial results can only be for 2D flow.  The starting point is comparing apples with apples - foil sections with foil sections.  You can speculate re a foil configuration; but you cannot engineer the structure until you have the section detail.  So initially you want to design viable (the best possible?) sections keeping the engineering constrains in mind - one of which is avoiding rotating the foil unit on each shunt, particularly as in order to keep weight down the unit may have to be lifted before rotating.

<<It is not easy to get a hydrofoil system with L/D better than 20 overall.>>

The moth L/D is around 15.