IRV has serious problems. It allows a sufficiently small minority of voters to safely register “protest” votes for minor-party candidates—but only as long as their candidate is sure to lose. As soon as their candidate threatens to actually win, they risk hurting their own cause by ranking their favorite first, just as they do under our current plurality system. IRV is therefore unlikely to be any more successful than plurality at solving the classic “lesser of two evils” problem.It’s straightforward to show, however, that this problem can only arise if your preferred candidate would be the loser in a Condorcet system. Hence, voting strategically yields the preferred Condorcet outcome.
To see how the argument works consider three candidates A,B,C and suppose that A’s supporters rank ABC, C’s supporters rank CBA, and that no candidate has an absolute majority. Then (regardless of the preferences of B’s supporters), B is the Condorcet winner. If B has the most or second-best supporters, then a (weakly) dominant strategy is for everyone to vote in line with their preferences, leading to B’s election. Suppose however that B has the smallest number of first-preference supporters and that A would win over C in a pairwise contest. Then, as stated in the critique above, the optimal choice for C’s supporters is to vote strategically for B, so that B finishes the first round ahead of C. The distribution of preferences then ensures that B is the winner. So, in this case, strategic voting does not produce the “lesser evil” as far as the majority of the electorate is concerned.
Things can tricker when there are n (greater than 3) candidates with a serious chance of winning. But the problems for Condorcet are even worse, since the method requires n(n-1) pairwise comparisons.
Taking this a bit further it seems likely that, whatever rule is chosen for resolving cycles, an implementable Condorcet system would be vulnerable to exploitation by strategic choices to run (or not run) particular candidates whose function would be to tip the balance in favor of some other candidate.
Taking this a bit further it seems likely that, whatever rule is chosen for resolving cycles, an implementable Condorcet system would be vulnerable to exploitation by strategic choices to run (or not run) particular candidates whose function would be to tip the balance in favor of some other candidate.
Possible solution in this case might be to impose stringent conditions for entry into the electoral race –either via some sort of financial yardstick (from individual donors) or some other indicator of a critical level of popular support not restricted to a narrow constituency; to weed out such candidates. Though this has problems in itself.
Then, as stated in the critique above, the optimal choice for C’s supporters is to vote strategically for B, so that B finishes the first round ahead of C.
But, because polling is inexact at best, it’s impossible to determine when to switch from CBA to BCA.
Personally, I’m a fan of approval voting, because of its simplicity. I understand that it (like all other systems) has its problems.
I recently participated in an election (for a small science fiction association). We did sequential (non-instant) run-offs, and the importance of non-primary choices became quite clear. In straw polls, candidates A and B had wide first-tier support. I asked for second choices, and candidate C swept the 2nd votes. Once everyone saw this, C ended up being elected.
The main problem with Condorcet is that it’s hard for the “average joe” to understand. I know these people have an argument about why it’s not, but any algorithm that requires college math isn’t going to appeal to a national populace.
I also tend to be wary of arguments like “Condorcet is the best because it is the only method which satisfies the Condorcet Criterion and four other criteria based on it.” The fact that they have an implementation in Python of all things scares me too.
I prefer approval voting to any of the other methods. My second preference is plurality, since it is simplest, but plurality wouldn’t work well in a country with a broader political spectrum than the U.S.
As noted in my previous post, Australians and Irish (not famously intellectual groups of voters) have no trouble with preferential voting systems, of which Condorcet is an example. But I agree with your scepticism about Condorcet.
I have no problem with preferential voting systems, though I think they introduce strategies for voters which are to some extent counter-intuitive.
It does bother me if a candidate wins because his or her supporters understand voting strategy better, rather than that candidate having the most popular support.
Dave, what’s wrong with Python?
(ooh, let’s have a programming language flame war to go with our voting system war ;)
I’m not set on Condorcet (I prefer the simplicity of approval voting). But I don’t think it’s hard to understand — I learned about it in 9th grade, IIRC.
I’d be more impressed with the criticism that STV rewards tactical voting if there were more actual examples adduced of tactical voting being rewarded. There have been tens of thousands of elections run under STV in Australia (just counting Federal, State and local elections) and there are no clear cases of this kind of tactical voting making a difference, and only a handful of cases where it is even plausible to claim tactical voting might have mattered.
Maybe it’s just indoctrination into the STV system, but when I see a case like the following, I think it’s natural that B should win.
48 votes for A then C then B
47 votes for B then C then A
5 votes for C then B then A
I was hoping Condorcet would come up when I posted my comments on the previous STV page about voting systems.
John Quiggen seems to prefer Proportional Representation to centrism. Well, the whole point of a Condorcet ballot is that a candidate has to appeal to multiple groups to elevate his preferences.
Borda can be rigged toward extremism by pushing the main opposition candidate toward the bottom. But there is no advantage to doing this in a Condorcet system.
Condorcet may seem complex, but it is exactly what pollsters do — they match up all candidates with each other in head-to-head contests. I don’t think this is too hard for voters to understand.
Approval is a reasonable alternative to Condorcet and also tends to choose the more centrist candidate. And it also can be polled for easily, so voters will have a better sense of the likely outcome if they want to change their votes.
Proportional Representation has its place, but it isn’t in a single winner election. You really want PR in an assembly.
The known problems with Condorcet occur when you get cyclic rankings. So you need a rule to decide which victories you want to include and which you don’t. Ranked Pairs / Maximize Affirmed Majorities is a reasonably understandable way to handle this. It isn’t hard to do. Pick the largest victory (by total number of votes), include the next largest victory, and leave out any victories that cause a cycle. You’re left with an unambiguous ranking.
Brian Weatherson’s example is interesting. The Condorcet victor would be C (beats B 53:47, beats A 52:48, B beats A 52:48). C is the second choice of a nearly divided set of 95 voters who really dislike the majority opposition candidate (A or B).
In STV, you get 47 happy B voters and 5 semi-satisfied C voters, but 48 very dissatisfied A voters.
If you set the last choice candidate to non-approved, the Approval winner would also be C, with 95 votes C to 52 B to 48 A.
So why should a candidate with lower approval win via STV? It’s only natural if you think using the logic of caucuses, where small losing blocks can sway the plurality when they switch allegiance.
À Gauche
Jeremy Alder
Amaravati
Anggarrgoon
Audhumlan Conspiracy
H.E. Baber
Philip Blosser
Paul Broderick
Matt Brown
Diana Buccafurni
Brandon Butler
Keith Burgess-Jackson
Certain Doubts
David Chalmers
Noam Chomsky
The Conservative Philosopher
Desert Landscapes
Denis Dutton
David Efird
Karl Elliott
David Estlund
Experimental Philosophy
Fake Barn County
Kai von Fintel
Russell Arben Fox
Garden of Forking Paths
Roger Gathman
Michael Green
Scott Hagaman
Helen Habermann
David Hildebrand
John Holbo
Christopher Grau
Jonathan Ichikawa
Tom Irish
Michelle Jenkins
Adam Kotsko
Barry Lam
Language Hat
Language Log
Christian Lee
Brian Leiter
Stephen Lenhart
Clayton Littlejohn
Roderick T. Long
Joshua Macy
Mad Grad
Jonathan Martin
Matthew McGrattan
Marc Moffett
Geoffrey Nunberg
Orange Philosophy
Philosophy Carnival
Philosophy, et cetera
Philosophy of Art
Douglas Portmore
Philosophy from the 617 (moribund)
Jeremy Pierce
Punishment Theory
Geoff Pynn
Timothy Quigley (moribund?)
Conor Roddy
Sappho's Breathing
Anders Schoubye
Wolfgang Schwartz
Scribo
Michael Sevel
Tom Stoneham (moribund)
Adam Swenson
Peter Suber
Eddie Thomas
Joe Ulatowski
Bruce Umbaugh
What is the name ...
Matt Weiner
Will Wilkinson
Jessica Wilson
Young Hegelian
Richard Zach
Psychology
Donyell Coleman
Deborah Frisch
Milt Rosenberg
Tom Stafford
Law
Ann Althouse
Stephen Bainbridge
Jack Balkin
Douglass A. Berman
Francesca Bignami
BlunkettWatch
Jack Bogdanski
Paul L. Caron
Conglomerate
Jeff Cooper
Disability Law
Displacement of Concepts
Wayne Eastman
Eric Fink
Victor Fleischer (on hiatus)
Peter Friedman
Michael Froomkin
Bernard Hibbitts
Walter Hutchens
InstaPundit
Andis Kaulins
Lawmeme
Edward Lee
Karl-Friedrich Lenz
Larry Lessig
Mirror of Justice
Eric Muller
Nathan Oman
Opinio Juris
John Palfrey
Ken Parish
Punishment Theory
Larry Ribstein
The Right Coast
D. Gordon Smith
Lawrence Solum
Peter Tillers
Transatlantic Assembly
Lawrence Velvel
David Wagner
Kim Weatherall
Yale Constitution Society
Tun Yin
History
Blogenspiel
Timothy Burke
Rebunk
Naomi Chana
Chapati Mystery
Cliopatria
Juan Cole
Cranky Professor
Greg Daly
James Davila
Sherman Dorn
Michael Drout
Frog in a Well
Frogs and Ravens
Early Modern Notes
Evan Garcia
George Mason History bloggers
Ghost in the Machine
Rebecca Goetz
Invisible Adjunct (inactive)
Jason Kuznicki
Konrad Mitchell Lawson
Danny Loss
Liberty and Power
Danny Loss
Ether MacAllum Stewart
Pam Mack
Heather Mathews
James Meadway
Medieval Studies
H.D. Miller
Caleb McDaniel
Marc Mulholland
Received Ideas
Renaissance Weblog
Nathaniel Robinson
Jacob Remes (moribund?)
Christopher Sheil
Red Ted
Time Travelling Is Easy
Brian Ulrich
Shana Worthen
Computers/media/communication
Lauren Andreacchi (moribund)
Eric Behrens
Joseph Bosco
Danah Boyd
David Brake
Collin Brooke
Maximilian Dornseif (moribund)
Jeff Erickson
Ed Felten
Lance Fortnow
Louise Ferguson
Anne Galloway
Jason Gallo
Josh Greenberg
Alex Halavais
Sariel Har-Peled
Tracy Kennedy
Tim Lambert
Liz Lawley
Michael O'Foghlu
Jose Luis Orihuela (moribund)
Alex Pang
Sebastian Paquet
Fernando Pereira
Pink Bunny of Battle
Ranting Professors
Jay Rosen
Ken Rufo
Douglas Rushkoff
Vika Safrin
Rob Schaap (Blogorrhoea)
Frank Schaap
Robert A. Stewart
Suresh Venkatasubramanian
Ray Trygstad
Jill Walker
Phil Windley
Siva Vaidahyanathan
Anthropology
Kerim Friedman
Alex Golub
Martijn de Koning
Nicholas Packwood
Geography
Stentor Danielson
Benjamin Heumann
Scott Whitlock
Education
Edward Bilodeau
Jenny D.
Richard Kahn
Progressive Teachers
Kelvin Thompson (defunct?)
Mark Byron
Business administration
Michael Watkins (moribund)
Literature, language, culture
Mike Arnzen
Brandon Barr
Michael Berube
The Blogora
Colin Brayton
John Bruce
Miriam Burstein
Chris Cagle
Jean Chu
Hans Coppens
Tyler Curtain
Cultural Revolution
Terry Dean
Joseph Duemer
Flaschenpost
Kathleen Fitzpatrick
Jonathan Goodwin
Rachael Groner
Alison Hale
Household Opera
Dennis Jerz
Jason Jones
Miriam Jones
Matthew Kirschenbaum
Steven Krause
Lilliputian Lilith
Catherine Liu
John Lovas
Gerald Lucas
Making Contact
Barry Mauer
Erin O'Connor
Print Culture
Clancy Ratcliff
Matthias Rip
A.G. Rud
Amardeep Singh
Steve Shaviro
Thanks ... Zombie
Vera Tobin
Chuck Tryon
University Diaries
Classics
Michael Hendry
David Meadows
Religion
AKM Adam
Ryan Overbey
Telford Work (moribund)
Library Science
Norma Bruce
Music
Kyle Gann
ionarts
Tim Rutherford-Johnson
Greg Sandow
Scott Spiegelberg
Biology/Medicine
Pradeep Atluri
Bloviator
Anthony Cox
Susan Ferrari (moribund)
Amy Greenwood
La Di Da
John M. Lynch
Charles Murtaugh (moribund)
Paul Z. Myers
Respectful of Otters
Josh Rosenau
Universal Acid
Amity Wilczek (moribund)
Theodore Wong (moribund)
Physics/Applied Physics
Trish Amuntrud
Sean Carroll
Jacques Distler
Stephen Hsu
Irascible Professor
Andrew Jaffe
Michael Nielsen
Chad Orzel
String Coffee Table
Math/Statistics
Dead Parrots
Andrew Gelman
Christopher Genovese
Moment, Linger on
Jason Rosenhouse
Vlorbik
Peter Woit
Complex Systems
Petter Holme
Luis Rocha
Cosma Shalizi
Bill Tozier
Chemistry
"Keneth Miles"
Engineering
Zack Amjal
Chris Hall
University Administration
Frank Admissions (moribund?)
Architecture/Urban development
City Comforts (urban planning)
Unfolio
Panchromatica
Earth Sciences
Our Take
Who Knows?
Bitch Ph.D.
Just Tenured
Playing School
Professor Goose
This Academic Life
Other sources of information
Arts and Letters Daily
Boston Review
Imprints
Political Theory Daily Review
Science and Technology Daily Review