A little bit of fluids 101: Two first-order things are happening on these aero bikes.
- Deeper tubes help keep flow attached, this is particularly evident at higher yaws
- Deeper tubes create more drag through both having more area and also creating a low-pressure region on the leeward side of the foil
This means that the fastest shape is an optimization problem between the shape (incl depth) and the yaw angle. You can make a foil too deep, which slows it down. Layering in optimization for weight, and it's easy to see why the rationale for all-around race bikes exists.
Similarly, wider forks are not always faster, as you are now balancing increased flow between wider blade spacing, with the depth/shape of those blades, and should also account for wheel/tire / spoke aero issues. It's an even more complicated optimization problem.
Where does this leave the 368? I would bet that real-world testing methods, either with an aerostick or using the zero-elevation/Chung method, would show that this bike is slower than an SL8. Might be closer in a tunnel, tho.