Based on some of the simulations, especially those using spatial models, the Supplemental Vote (SV) used in London may well be worse than FPP in terms of both social utility and Condorcet frequency. At least with FPP it is easier for voters to compensate using strategy. It occurs to me that SV could be improved by allowing multiple second choices, much like Approval. Condorcet efficiency would probably still be slightly worse than for traditional runoff, but social utility might be better, since people have incentive to vote for multiple "close" second choices, but not for "distant" seconds. It would also be a better approximation of traditional runoff than SV (which only allows one first choice and one second choice), since at the time you vote your first choice in traditional runoff, you could have many potential second choices. Not an ideal system by any means, but might offer a transition toward better methods (both ranked and unranked), since it has elements of both. SV as practiced has the potential to backfire, with a bad outcome turning voters off from further experimentation. http://www.electoral-reform.org.uk/new%20separate%20pages/leaflondon.htm
