It's taken me a while to fully understand the D3 dynamics in an R22, but Frank Robinson's explanation (on Dave's site) does make sense. In a nutshell:
1. Rotor coning and inflow cross-coupling compensation.
2. Pitch and roll velocity cross-coupling compensation (wee-wa)
3. Vertical component gust reduction - more as a side effect.
4. Increase in revolutions to respond to cyclic input.
Point 4 could be seen as a problem, but since a teetering rotor has acceleration control for cyclic input until equilibrium establishes itself at a fixed pitch or roll rate, then this slow down in response probably makes the machine more flyable (once you learn not to overcontrol). The final proof is that it flies in the direction you point it.
Mart