I don't claim that this wording is in its best final form for
a public proposal. But it's my suggestion so far:
A beats B if more voters have ranked A over B than vice-versa.
If 1 alternative beats each one of the others then it wins.
But, in any case, if there's a group of alternatives that all
beat everything not in that group, then the winner must be
chosen from that group.
The group of alternatives that the previous paragraph narrows
the choice down to shall be chosen from by Condorcet's "votes-against"
The winner is the alternative least beaten by any other
alternative that beats it, as measured by how many voters rank
that other alternative over it.
[Alternatively: The winner is the alternative over which fewest
voters have ranked the alternative ranked over it by the most
voters. I like that, but the wording could be accused of sounding
And yes, as I said, because of people's liking for brevity, I
don't like having to lengthen Condorcet's definition by including
the Smith Criterion. But the academics may force us to do that.
As I said, I'd prefer just leaving the Smith Criterion out, and
then adding it as necessary, if & when criticism based on
candidate-counting criteria turn people against plain Condorcet.
At that time people will understand the reason for the Smith
Criterion--since the candidate-counting criticism will have
convinced them of the need for it.
Besides, as I said, we could also just propose plain Condorcet,
with a special rule to disqualify Condorcet losers. An obvious
patch? Sure, but, as I said, it's natural that a need to meet
more than 1 standard can require more than 1 rule. Besides, as
I also said: If only such a simple patch could do something about
Copeland's failures in regards to majority rule, defensive strategy
I emphasize that, even if we have to propose Smith//Condorcet,
instead of plain Condorcet, it isn't necessary to define the
Smith set--it's sufficient to briefly & simply state the
Smith Criterion, as I did above.