To Martin on SFC

[MIKE OSSIPOFF]

I've had to be away from the computer for a while. I probably won't
be able to reply as often as I have been, but I'm going to try to soon
reply to the messages that I've meant to reply to.

Martin wrote:

Let's compare this to the Independance from Clones Criterion. Blake's
defn of this is the following (with appropriate defns of clone set,
eliminated, etc):

 >>>>> "If there are alternatives X1, X2 ... Xn that are a clone set,
and if one of these clones is eliminated from every ballot, then, if the
winner for the old ballots was in the clone set, the winner for the new
ballots must also be in the clone set. If an alternative outside the
clone set won for the old ballots, the same alternative must win for the
new ballots."

I reply:

Like the other definitions at Blake's website, that doesn't seem a
good one.

Of course it isn't complete without a definition of clones, but,
just from a brief initial glance, it looks as if Plurality passes
the criterion as written above.

I'd rather define it in terms of sincere preferences, and with a
stipulation of sincere voting.

Martin continues:

Now, we could rephrase this using your type of wording, and the first
clause would be replaced with:

"If there are alternatives X1, X2 ... Xn that are a clone set, or are
close enough to a clone set, and..."

I reply:

But you know, don't you, that that isn't how I define SFC. My definition
stipulates that no one falsifies.

Martin continues:

But what does "close enough" mean? Well, exactly the same as your "on a
scale sufficient": it might mean a single voter who votes them
non-adjacently, or it might mean lots of voters voting them non-adjacently.

I reply:

But haven't we covered those questions? I replied to them (on May 9th,
wasn't it?).

Martin continues:

Here's another example: "IRV elects the Condorcet winner, provided that
third parties do not occur on a scale sufficient to change the outcome."
- an absolutely true statement. But if you check it, all it's actually
saying is that IRV elects the Condorcet winner, provided there are only
two parties in the election.

Now, you claim that for the wv methods you propose, the typical levels
of falsification will be typically unable to stop SFC's gaurantees from
working. That's all very well and good - but it's just a claim, and it
is entirely possible for a method to pass SFC, but to fail this further
claim.

So what I'd really like is to have some evidence for your claim. I'd
also like to see both instances of "typically" given values: for
example, in terms like "if the CW beats B by X%, then falsification can
occur on a scale of up to Y%, and B will still not be elected."
(provided the rest of the conditions hold.

I reply:

In our EM polls, there have been short rankings, but not much 
order-reversal. Since you bring it up, I'll say that it looks as if Blake
perpetrated order-reversal in the voting systems poll. But surely it's
obvious that that it is unlikely that the outcome would be changed
by order-reversal. I was overly concerned about that. That's why I
designated Manual instead of Cloneproof SSD.


Martin continues:

Incidentally, you criticise my example methods somewhat. The point here
is that they were examples. I like to keep my examples simple. I thought
it was fairly obvious that they weren't serious proposals, given how
they involved oracles, devils, and similar such things.

More later.

Mike Ossipoff


_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com