[EM] The Hippopotamus Logic

Mike Ossipoff dfb at bbs.cruzio.com
Sat Nov 9 16:12:00 PST 1996

donald at mich.com writes:

I'm replying to Donald's comments a little way down this letter.
I'd delete the rest if I could delete blocs of text. Anyway,
the part I'm replying to is a short distance down in this letter.

> Dear List,
> DEMOREP1 wrote:>
> >If there was a yes/no vote on the acceptability of each candidate, the
> >results might be (assuming that the second choices are "true")
> >A 46 yes, 54 no- defeated
> >B 100 yes, 0 no- survives
> >C 34 yes, 66 no- defeated
> Donald: I question what you are doing here. You have no right to use only
> the first and second selections as yes/no votes. The third selections that
> 80 percent of the voters chose to make are also yes votes. The results
> should be:
>           A   80 yes  20 no
>           B  100 yes   0 no
>           C   80 yes  20 no
> DEMOREP1 wrote:>
> >A yes/no vote in Donald's third example would produce
> >(assuming 2nd choices to be true)---
> >40A      40
> >2AB       2      2
> >4AC       4              4
> >20B             20
> >30C                     30
> >2CA       2              2
> >2CB              2       2
>           ---    ---     ---
> >Totals   48A    24B     38C
> >No majority acceptable candidate
> Donald: You are using Approval Voting here - this suggests to me that
> Approval Voting is not a valid election method - at times it will not come
> up with a majority winner. I do not approve of Approval Voting.

Approval isn't as good as the best rank-balloting methods, such as
Condorcet & Smith//Condorcet, but Approval is better than the worst
rank-balloting methods, such as Instant-Runoff. Approval is surprisingly
good, for such a simple method, though it can't be as good as the best.

>      I would approve a yes/no vote on the first selection with the loser
> being dropped. Using Steve's example:
>         >46 ABC          A 46 yes  54 no
>         >20 B            B 20 yes  80 no
>         >34 CBA          C 34 yes  66 no
>      Candidate B has the largest no vote and is dropped.
>      Mike and Steve, the two "big-name Condorcet popularists on this EM
> list", both claim to be champions of majority winner but they gloss over
> the fact that eighty(80)percent of the voters in this example voted for
> candidates other than candidate B. If candidate B is so great why did he
> not receive more votes. I would say that candidate B is most likely an
> adequate person to serve as a caretaker president - not a superstar.

You're confused about what majority wishes means. The fact that 80
percent didn't vote B as their favorite is irrelevant to the
fact that a majority indicated preference for B over A, and a majority
indicated preference for B over C. Those are clear-cut majority

You, judging by your list above, have assumed that the 80 who didn't
vote B in 1st place have voted "No" on B. Nonsense. They haven't
voted that B shouldn't win. They've indicated that they'd rather
have B than their last choice.

Can someone say that 80% not voting B in 1st place is a majority
statement? Sure, it means that a majority don't consider B to
be their 1st choice. But the election is about choosing 1 alternative
instead of others, and you can't (or shouldn't) ignore it when a
majority say that they'd rather have B than A. I repeat that 
no majority has voted that B shouldn't win. You see, you'd ignore
what a majority did say, and hear something that they didn't say.


> Donald,
> .-


More information about the Election-Methods mailing list