As Blake said, the attempts that had been made so far, at the time of his statement, to write a precise criterion based on the 1st Choice standard proposed by David Marsay were unsuccessful. I don't deny it. But Blake seemed a little too optimistic when he said that it was clear to him that a precise 1st Choice Criterion would never be defined. Plainly I can't claim that my several unsuccessful attempts at that mean that I'm good at writing difficult criteria. Ordinary GMC is useful, but it's also quite simple, unlike what one gets into when trying to write a 1st Choice Criterion. The standard, as expressed by David is so obvious, and the importance of not having to completely abandon our favorites is so obvious & well-understood, that I greatly underestimated the difficulty of writing a 1CC, so much so that several times I mistakenly believed that I'd written a usable one when I hadn't. But with or without a precise criterion for it, the 1st Choice standard's importance is clear. I've shown examples of Margins making people have to abandon their favorite. Is a criterion needed to prove that that's udesirable. I've been the only person pursuing the effort to write a 1CC, and maybe that gives Blake confidence that none will be written; but even if I haven't demonstrated myself to be especially skilled at difficult criteria, don't celebrate prematurely, Blake. The underlying standard is so important that the effort will continue. There isn't really any question,is there, that, with better attention to detail, without the mistaken belief in the simplicity of the task, such a criterion can be written. Anyway, in the meantime, the criterion that I'd been calling "LO2E", and one similar to it, will do fine. Earlier Blake found a fault with LO2E, and I no longer have that letter stored. I'm not being careless, and haven't found a logical fault, but of course there could still be one. These criteria will serve the purpose that 1CC was intended for, putting certain methods' failings on a more precisely-described footing (hopefully). I'm a little concerned that, since I can't find Blakes letter about LO2E-2, that these criteria might have a problem that I haven't found, but I'd like to post them anyway, here, and in a separate message: I'm writing them both in the weak form. The strong form is gotten by replacing "over a more-liked alternative" with "equal to or over a more-liked alternative". A simple defensive strategy criterion: A group consisting of a majority of all the voters, who all vote a certain particular candidate A over a certain particular candidate B, should have a way of voting that will ensure that B can't win. Accomplishing that shoudn't require any members of that group to vote a less-liked alternative over their favorite. LO2E: Replace the last sentence with "Accomplishing that shouldn't require any members of that group to vote A over a more-liked alternative. That change was added in order to lower the bar enough for Condorcet(EM) to strictly meet the criterion, instead of just to meet it for all practical purposes. Mike Ossipoff
