<p dir="ltr">I'm changing NEO's dfn in one regard:</p>
<p dir="ltr">Plain NEO:</p>
<p dir="ltr">If 1 candidate wins at Approval Nash equilibrium, s/he wins.</p>
<p dir="ltr">If more than 1 do, then re-apply the above paragraph to them, with only them in the count.</p>
<p dir="ltr">Repeat till there's only 1.</p>
<p dir="ltr">If, at any stage there is no Nash Equilibrium, or everyone wins at Nash equilibrium, the most top-ranked candidate wins.</p>
<p dir="ltr">I defined a cohort as a set of voters who prefer and vote the same. My Nash equilibrium definition uses that.</p>
<p dir="ltr">I define a "combo" as a set of voters who share at least 1 pairwise preference.</p>
<p dir="ltr">Cohort or combo could be used. When I say NEO, I refer to the cohort version.</p>
<p dir="ltr">In the situations of interest (for the properties I want), the only difference between cohort & combo is that, in the chicken dilemma situation, cohort elects C, and combo elects A.</p>
<p dir="ltr">People at EM prefer C. That's an advantage for cohort NEO.</p>
<p dir="ltr">I don't know of a reason to use combo NEO instead of Smith//MMPO, since both elect A.</p>
<p dir="ltr">Michael Ossipoff</p>