[EM] Condorcet at the Instantrunoff Mailing List

[Markus Schulze]

> This message was sent to the original instantrunoff list by jquinn+irv
> but he consented to have it forwarded to the freewheelin' list.  So,
> when he talks about "this list" he means instantrunoff classic.
> 
> 
> >From jquinn+irv:
>
> So far, this list's "last word" on Condorcet is that it is not a valid
> topic for discussion here. Since the volume is somewhat down again
> (last digest only contained one message), I think now is an
> appropriate time to deal with this issue.
>
> [[[DISCLAIMER: In this message, I will talk first about censorship on
> this list, then explain one currently relevant advantage of Condorcet
> voting over IRV voting. Please understand that I do not mean to open a
> flame war on either point. I have respect for the CVD, this list's
> administrators, and the system of IRV; I consider myself an ally to
> all three. I also know that the fact that truly off-topic messages are
> censored is the reason to have a moderated list in the first place.
> Therefore, I do not mean to use the word "censorship" in an
> inflammatory sense, but merely as an acknowledgement of the
> seriousness of the choices involved.
> END DISCLAIMER]]]
>
> I understand that this list's primary purpose is to organize for
> activism around IRV. However, it cannot help but serve also as a means
> of self-education, both on the details and ramifications of IRV, and
> on the mechanics of activism and mailing lists. It would be very
> unhealthy to restrict the scope of discussion to only pro-IRV
> messages. Of course, it would also defeat the purpose of this list if
> rancorous debate between voting systems drowned out the focused
> activism. Therefore, I recommend a policy of partial censorship. Allow
> some percentage (say 15-25%) of list traffic to be devoted to
> discussion of voting systems, and only censor the messages above that
> limit.
>
> There is all the difference in the world between partial censorship
> and a blackout. I have heard from at least 4 different voting system
> enthusiasts that they have been offended by what they perceive as the
> CVD's policy of simply not discussing alternatives to IRV. One person
> was so angered he even stooped to name-calling. I was pointing out
> advantages of IRV over approval voting and he said "that's what the
> CVD irvies all say." It is one thing to try to keep things simple when
> writing editorials for general consumption; but on this list, among
> people who are already well-versed in the advantages of IRV over
> plurality voting, discussion _must_ be more open or you will isolate
> and fractionate yourself more than any flamewar could. To this
> purpose, I suggest that a CVD member open the discussion with a
> defense of IRV as superior (either theoretically or practically) to
> other single-winner voting systems.
>
> Finally, I want to point out one very topical advantage of Condorcet
> (aka Instant Pairwise Runoff Voting/IPRV, aka Instant Round Robin
> Voting/IRRV) over IRV: it is easier to count and less vulnerable to
> fraud. If, to take one not-so-random example, I marked a '1' next to
> both Gore and Buchanan, that would be a spoiled ballot in an IRV
> presidential election.  With IPRV, the vote would be spoiled in the
> Gore vs. Buchanan pairwise race, but would still count in the Gore vs.
> Bush race where it mattered. Moreover, since the election will never
> in a thousand years hinge this closely on two separate pairwise races,
> any hand recount can focus on one race and each precinct can just
> report 2 recount numbers (for the two candidates in the dead heat)
> which add up nicely to the statewide total. With IRV, since the whole
> election can hinge on who is eliminated in which order, complete
> preference data for every ballot must be transmitted to one central
> location for counting - a process much more vulnerable to large-scale
> fraud (and of course just as vulnerable to small-scale fraud).
>
> In other words, if you're arguing for IRV and someone says, "That's
> too complicated when Palm Beach can't seem to vote or count properly
> with today's system", you can say, "Well, a close relative of IRV
> called IPRV is actually less vulnerable to misvotes and miscounts than
> what we've got."