Clarifications re: SD Algorithm

[Mike Ositoff]

If you sequentially drop weakest defeats, without regard to whether
they conflict with stronger ones, till something is unbeaten
that beats something else, then, in a public election where there
are no pairwise ties, and where therefore anything unbeaten beats
something else, isn't that going to give the same result as
plain Condorcet(EM)?

I hadn't thought of how it would be programmed, or written
as a precise algorithm. Checking for cycles sounds like quite
a programming task. Is it one of those things that becomes
un-computable when the number of candidates becomes large,
like the traveling salesman mileage-minimization problem when
the number of cities becomes large?

Mike