The diagram below nicely illustrates what happens as speed builds.
The plot takes a constant true wind angle (heading relative to true wind direction). As boatspeed increases, the component of apparent wind from dead-ahead gets bigger, while the true wind velocity stays constant. Therefore the wind we feel on our sail moves progressively further forward.
 |
| Diagram taken from the interesting article here: http://boards.co.uk/how-to/how-fast-can-we-go-the-science-of-speed.html#AmZd7dgCcvAEmL31.97 |
As apparent wind direction moves forward, we have to sheet on in order to keep an angle of incidence between sail and apparent wind. But moving the boom closer to the centreline increases the component of sail force that wants to heel the boat over. For the same sail force, drive decreases and overturning moment increases. Hence we need righting moment to resist the added heeling moment.
 |
| On the left we have the boom eased and the sail twisted. Sail force (orange) is mostly directed forward. On the right the boom is near centreline and the leech is tightened by sheeting on. In these diagrams sail force is identical but the component to leeward (responsible for heeling moment) is much larger when the apparent wind rotates forward. In reality, when the apparent moves forward, it also increases in speed. So sail force would be larger, exacerbating the problem. |
Recent work with AC45 Turbo test platforms, and smaller cats, has shown that differential rudder rake, such that the windward elevator foil produces downforce, can add significant speed.
It may be that using the windward foil to pull down could pay, bringing about a return to four-point systems.
But it is certain that span restrictions that require multiple surfaces to get sufficient lift on the leeward side are driving development to a dead-end. If such restrictions persist, more changes are inevitable as people invent elaborate rule-cheats such as having several foils in each hull.
Much better to encourage simplicity.