and thank you everybody for fuelling the discussion with sound arguments.
My purpose was not to focuse on a potential loophole of rule 8, instead it's about measuring the "gap" between a theorical optimum solution (mimimun drag @ a design speed) and the best existing solution.
My theorical optimum solution would be (if computations are not stupid) a surface piercing foil at 67° from vertical, with a 56cm span.
If the board with no load has a 45° angle, and if it can flex under full load at 67° then we have already the optimal solution which meets perfectly rules 8.
In other words, with the flex, insertion from below is no more necessary to get the optimum.
That's all I wanted to enlight, I m not proficient with measurement or fair interpretation of the rules, so I stick to what I believe I understand better.