# NOTA in a circular tie

Steve Eppley seppley at alumni.caltech.edu
Tue Dec 17 10:27:52 PST 1996

```>Mr. Eppley wrote:
>>But it would be nice if he [Demorep1] would explain WHY he thinks
>>B's disapproval is more important to the voters than the fact that
>>a majority prefer B more than A.

Demorep replied:
>I wrote earlier but repeat- a relative vote (B > A) based on an
>absolute disapproval of the candidate (B is unacceptable to be
>elected) is a fraudulent vote (and gives any such B candidate who
>might otherwise be the relative winner a false sense of public
>approval).

I think Demorep should reconsider the meaning of a circular tie
involving NOTA, as well as the meaning of "absolute".  Approval
or disapproval is really a relative pairwise comparison between a
candidate and NOTA.  An absolute would specify *how much* better
or worse than NOTA is the candidate (as in a rating system such
as -10 to +10).

I see nothing fraudulent about picking the choice which wins the
normal circular tie breaker, giving no special treatment to the
NOTA choice.

>>The question involves a circular tie between A, B, and NOTA:
>>     A > NOTA > B > A
>>Why should NOTA be treated differently than a normal choice in a
>>circular tie?

>A vote for NOTA (none of the above) (if it is going to be on
>ballots) is a special case and is an absolute vote, not a relative
>vote.  That is, if a majority of the voters vote NOTA, NOTA wins
>(i.e. NOTA goes to the head of a voter's ballot and the majority's
>combined ballots making NOTA a de facto NOTB (none of the below)).

I don't know what Demorep means by "a majority vote for NOTA" nor by
"NOTA goes to the head of a ballot".  Maybe he means a majority
who voted NOTA as their first choice?  Why is that relevant in a
pairwise method?  And what does he think it means when a voter ranks
NOTA after a first choice but ahead of some choices?--it looks like
he suggests ignoring any NOTA which isn't first choice on a ballot.

>I would suggest not having NOTA on the ballot-- i.e. require a
>voter to approve or disapprove each candidate individually.

when Disapproval is combined with Condorcet, even when forced to use
disapproval transitively (i.e., NOTA in rankings).  For example:

Sincere:                        Order-Reversal:
46: AB/                         46: A/B           <-- reversal
20: BA/                         20: BA/
34: /BA                         34: /BA
B would beat A.                 B beats A.
B would beat NOTA.              NOTA beats B.
B would be elected.             A is elected.  Reversal succeeds.

I think the tactical opportunities are even worse when the voter can
approve or disapprove candidates intransitively.

I don't recall any attempts by Demorep to address any of the tactical
concerns of the method he has been espousing.

It would also be more tedious for voters to have to specify approval
for many candidates instead of just ranking the NOTA/NOTB token as if
it were a candidate.

If Demorep has explained WHY he suggests individual approval or
disapproval instead of ranking NOTA/NOTB, I've missed it.  This
appears to be just another opinion with no attempt to justify it.

>A reasonable ballot access requirement, (such as 0.5 percent of the
>votes cast at the last election for the office) combined with a
>majority approval or disapproval vote (especially the latter)
>will eliminate the extremists for executive and judicial offices.
>I repeat- leftwing and rightwing extremists can run for a p.r.
>legislative body and expound their views as in sundry European p.r.
>elections.

Yes, nations survive the freedom of speech given to a few elected
legislators who have little popularity.  But in single-winner
elections, majority disapproval seems to be an overkill device
rendered unnecessary by Condorcet when there's at least one good
candidate competing.

What's the meaning of "N% of votes cast" when voters cast ranked
ballots?  Is this another reference to first choices?

---Steve     (Steve Eppley    seppley at alumni.caltech.edu)

```