[EM] IRRV as best popular name for Condorcet voting

Steve Eppley seppley at alumni.caltech.edu
Sun Sep 19 12:53:09 PDT 2004


Hi,

Paul K wrote about my suggestion to drop the V from IRRV:
> I had something of the same notion (and being more nearly 
> an average voter than a voting methods scholar, I agree
> with this). 
> 
> But... for the same reason there shouldn't be a "V" on 
> the end of IRV. It should be IRO for instant run off. 

Years ago, some people did refer to it as IRO.  But runoff
is the correct term; it's not two words.  Despite that,
I agree IRO is a better name than IRV.  IR would be too 
short, probably, since IR already means infrared and 
perhaps dozens of other things too.  

> If there's a V on the end of IRV, we need a V on 
> the end of IRRV to keep FROM confusing everybody. 

I apologize if I missed an earlier argument that justified
that opinion, but I don't understand why I should believe
that's correct.  To keep from confusing people, the title
or intro of a document or letter could refer to it as
the Instant Round Robin voting method, and then abbreviate
it as IRR.

> This was why earlier I argued that we should keep "voting" 
> and "vote-counting" methods separate. I "vote" the same
> way whether its IRV or IRRV - the difference is in the
> counting, and that's what we're already suspicious
> about... 

In a technical sense, yes, we would vote the same way-- 
assuming IRV permits candidates to be ranked as equally 
preferred, which we IRR advocates would permit for 
both IRR and IRV.

But the candidates to be ranked would likely be very 
different.  Under IRV, I expect we'd still see what we 
see today: two big parties that each nominate one candidate 
apiece--requiring primary elections to winnow their 
contenders down to one--plus a few third-party candidates 
who are sure to lose, and the two big candidates will 
continue to be polarized by their need to get off-center
votes to be nominated and their need to stay fairly 
true to that off-center base in the general election 
(to get their base to show up to vote on election day).
If a candidate were reckless enough to position himself 
in the middle, he'd be sandwiched by surrounding 
candidates and lose and appear unpopular given IRV's way 
of measuring popularity. (Like John McCain, Condorcet 
winner in 2000, would have appeared relative to Bush 
and Gore, if the voters' preference orders in 2000 had 
been tallied by IRV.)

Under IRR, on the other hand, parties would not have 
a strong incentive to nominate only one candidate, 
and there'd be strong incentives to nominate more than 
one: they could increase turnout of their supporters 
on election day by nominating a diversity, they could 
avoid the fratricide of primary fights, and they could
avoid putting all their eggs in one basket.  To be 
successful, candidates would need to adopt centrist 
compromise positions.  This would allow voters to rank 
less-corrupt centrists over more-corrupt centrists,
electing less-corrupt centrists.  I couldn't ask much 
more of a voting method, since elections are such crude 
instruments.  Democracy is not about being fair to each 
voter, as one member of this list recently asserted 
during our discussion of the electoral college; it's 
about aligning the incentives of society's leaders 
with the well-being of the people.

--Steve




More information about the Election-Methods mailing list