Margins -- Blake is right -- slight amplification (fwd)

[Mike Ositoff]

> 
> On Sun, 27 Sep 1998 20:24:19   Mike Ositoff wrote:
> >> > Hugh R. Tobin wrote:
> >> > 
> >> > > I agree that votes-against does not permit a rational "defensive"
> >> > > strategy of truncation; moreover I remain to be convinced that in
> >> 
> >> Not sure what you mean by that; defensive truncation, in VA, makes
> >> it impossible to profit by voting an insincere preference, and
> >> penalizes order-reversal.
> However, if you guess wrong, and you are not the CW, it lowers your
> chance of winning.  It never increases your chance of winning.  The

With any method you lose by guessing wrong. Let's not declare that
to be a special problem of VA :-)

> order reversal strategy, however, can actually defeat an order-reversal
> campaign by saving your candidate, as opposed to getting a third 
> candidate elected.  For this reason I do not consider the truncation defensive strategy to be rational, and therefore a more effective
> deterrent than order-reversal.

True enough: VA's defensive truncation doesn't make the CW win. It
merely prevents to order-reversal from happening in the first place,
by deterring it.

And, as I said, Margins' anti-order-reversal strategies actually
make order-reversal safer. That's true, in the A,B,C examples,
if C uses defensive strategy, and it's true if B voters use
the defensive strategy of declaring that they're going to rank
C over A. By so doing, they're inviting A voters to order-reverse
against B.

I don't know if you understand the problem of an anti-order-reversal
strategy that prevents one side from successfully order-reversing,
while laying out a red carpet to invite the other side to order-reverse.


> >> 
> >> 
> >> > > margins  it would ever be rational for a voter to truncate, but not to
> >> > > reverse order.  
> >> > 
> >> > It may not have been clear enough that the second point relates to a
> >> > separate issue from the first, that is, a charge made on this list that
> >> > the margins tiebreak allows an "offensive" truncation strategy, which
> >> > fails in "votes against".  Such a strategy has no independent
> >> > significance because it is only a weaker alternative to order-reversal
> >> > strategy, which in the right conditions can work under both methods.
> >> 
> >> No, truncation has independent significance because it occurs in
> >> all rank-balloting elections, and will surely be common in public
> >> elections. Margins is vulnerable to it, and VA isn't.
> All voters who know how Votes-Against works will use the random-fill
> strategy instead of sincerely or insincerely leaving candidates unranked.
> 
> Of course, if a system penalizes voters for leaving candidates unranked,
> doing so will not play a role in successful strategy.  However, those
> voters who catch on to random-fill will not be affected by the penalty.
> In aggregate, there votes will be scored as if they were using Margins.

In a recent series of letters, I told why insincere extension
won't be a real problem with VA.

> >> 
> >> To say that truncation isn't significant because the would-be
> >> truncator could order-reverse instead in VA is like saying that
> >> door-locks are useless because the burglar could instead saw a
> >> hole in the door. It's more trouble, and it carries a great risk.
> If you prefer B to C, but truncate, you do so because you hope to
> create a large win against B.  The more effective way to do this
> is to actually rank C over B.  Of course, this goal may turn out to
> be undesirable, if it makes C win over all.  However it makes no
> sense to only half pursue defeating B because you are not sure
> if it is desirable to do so or not.
> 
> 
> 
> -----== Sent via Deja News, The Discussion Network ==-----
> http://www.dejanews.com/  Easy access to 50,000+ discussion forums
> 
>