Hello. This email is an invitation to join
the range voting email group and save the world.
Range voting is an improved voting
method related to approval voting. It will vastly
improve those governments which adopt it. The improvement is huge
and the cost is near zero. Many defects of the US
Can anybody tell me whther any interesting voting methods are mentioned in the
Bible?
Election-methods mailing list - see http://electorama.com/em for list info
it (for the moment) is located at
http://math.temple.edu/~wds/crv/RangeVoting.html
please check it out. Any help in constructing and improving this web page
will be appreciated. You can add new content by emailing me files.
You can improve the presentation of old content by grabbing, editing
Well, according to that definition, Range Voting is a Condorcet
method, since if you erase all candidates and all numerical votes for
them in all range votes - except for two candidates A and B - then A
will beat B in the resulting 2-choice election if and only if he beat
B in the original
robla:
My point was Condorcet only considered ,,= relations WITHOUT numerical
magnitudes,
just yes of no, as votes, hence the distinction between my 1st + 2nd
defintions of the Condorcet concept, dd not occur in his mind.
That would be /the/ definition of the Condorcet winner. I sincerely
Robla:
Breaking up the definition into two sections as though both are
completely accurate and equal definitions of the Condorcet winner
criterion makes it appear as though the Condorcet winner criterion is
poorly defined. Do you believe that the Condorcet winner criterion is
poorly defined?
As someone who has a rather bad habit of calling simple and honest
mistakes lying, I can sympathize with your reaction. Either way,
though, range voting isn't Condorcet compliant, as Condorcet is defined
by the concept of pairwise victories, which do simplify to the mere
greater than or less than
Robla: Rob Lanphier wrote:
Hi Warren,
I'm interested in Range Voting, since it appears to be popular among
many electoral reform advocates here.
I too get the impression that it is generally considered among the
best voting method possible, BUT for one truly fatal flaw. It
suffers
If [the Condorcet winner criterion] means (2a) if we redid the same
election but using a DIFFERENT election system, namely majority vote,
and demanding votes `logically consistent' with the originally-cast
votes, then A would win
This is what it means. I'm quite confident that the
. Everyone else, give them the lowest
possible ranking. No information about what other voters are doing is
needed.
REPLY BY WARREN SMITH:
I was right and you are wrong. Here is
a counterexample. Let the other voters create a tie for first among A and B,
*or* a tie among B and C, you do not know which
WDS: ...
If [the Condorcet winner criterion] means (2a) if we redid the same
election but using a DIFFERENT election system, namely majority vote,
and demanding votes `logically consistent' with the originally-cast
votes, then A would win
Robla: This is what it means. I'm quite confident
Adam Tarr: I'll now respond to Warren's earlier message.
I don't debate that the more-favored front runner first, less-favored
front runner last strategy is useful (often optimal) in Borda, but I
can't easily imagine a scenario where it is useful in Condorcet.
WDS: I did not say it was the best
1. condorcet.org definitions page:
Name: Condorcet Criterion
Application: Ranked Ballots
Definition:
If an alternative pairwise beats every other alternative, this alternative must
win the election.
Pass: Black, Borda-Elimination, Dodgson, Kemeny-Young, Minmax, Nanson
(original),
On the probability that insincerely ranking the two frontrunners max and min, is
optimal voter-strategy in a Condorcet election.
--Warren D. Smith Aug 2005--
MATHEMATICAL MODEL: 3-candidate V-voter Condorcet elections
with random voters (all
On the probability that insincerely ranking the two frontrunners max and min, is
optimal voter-strategy in an IRV (Instant Runoff Votng) election.
--Warren D. Smith Aug 2005--
MATHEMATICAL MODEL: 3-candidate V-voter IRV elections
with random
OK, I can see I'm hitting a wall of opposition here. This whole issue is
a red herring (i.e. distraction from my main point) so let us not be
too distracted by it. The central issue which we had started from is the
question of which is better - range voting or Condorcet methods?
So instead of
It was recently claimed on EM that condorcet had simpler rules than
range. I dispute that. I challenge people to write computer
programs to perform condorcet and range elections. I have so
far never encountered anybody who produced a shorter program for condorcet.
Not even close.
For any
In some sense the range versus Condorcet debate is a red herring
since Condorcet methods have, I think, no chance of actual adoption
by governments. And range does have a chance. So for practical purposes,
forget Condorcet. Why do I say that?
Well, in our real-world-voter study of range
I disagree with the claim they are. Democracy is about
choice by the voters. 2 choices is not enough choice.
Furthermore in the contemporary USA, 98% of the time
incumbents are re-elected if try. So really it is a 1-choice
system. Congressmen are more likely to
die in office than to lose an
In fact let me elaborate. Although my critics claim it
is not clear I have really shown Condorcet methods must lead
to 2-party domination (I think it is clear, except I admit
that the winning-votes + equalities-permitted enhancements of condorcet
seem to permit Condorcet to perhaps escape from
But if you do consider 2-party system to be democracy, then
if you like 2-party domnation state, why bother to consider any
voting system other than plurality at all?
I mean, plurality (1) maximally leads to the situation (2 parties)
you like, and (2) if there are only 2 candidates, then
robla: Condorcet has zero chance in 2005. It has a small chance in 2010, and
better than even odds in 2050. That's assuming we ignore your advice
and actually continue our work.
--what is your strategic plan? One can make statistical estimates of chances
based on polls and one can estimate
Rob Lanphier re the Center for Range Voting:
If you had the kind of backing that CVD has, I might believe you.
However, in terms of popular voting reforms, only CVD can make the claim
that they've got the political organization and the momentum to follow
through right now. CAV/AAV is making
Dave K:
Range voting is very robustly the best among about 30 systems tried including
a couple condorcet systems according to my giant
comparative Bayesian regret study in 2000. OK, maybe you can attack that.
Maybe you can say I did not put in your favorite system or favorite
voting
As was recently pointed out, it is correct that with range ballots
run on ordinary plurality voting machines, slots (e.g. levers on
NY-style machines) get chewed up 10 times faster than
with plain plurality voting. Assuming 10 levels.
With L levels, L times faster.
Consequently if enough
Scott Ritchie:
Indicating a ranked ballot on a machine not designed for it is no more
difficult than indicating a ranged ballot. This follows naturally from
the fact that you can do a one-way transformation on a ranged ballot to
a ranked ballot.
There's a great picture of an old New York lever
range and borda really alike? not.
Scott Ritchie:
Range voting is just a Borda count with a bunch of throwaway candidates
--WDS REPLY:
Range voting and Borda indeed have a lot of similarities,
but they also have some extremely crucial differences. In fact I would
say that range voting keeps
range with L levels chew up L slots? - other ideas by Lomax me
It is actually possible to do range with L levels on plurality voting machines
with
only chewing up log[base2](L) slots per candidate, not L.
This is a huge improvement: L=512 becomes 9.
Unfortunately, it requires the voters to
I believe the brute force approach of just solving the NP-hard redistricting
problem
perfectly, is not feasible. There are probably ten-thousands of census blocks
and exponential runtimes with that much input just do not happen, even with
all the computer power on the planet on your side.
The graph partitioning problem is NP-complete:
problem ND14 in Garey Johnson: Comnputers and Intractibility, a guide
to the theory of NP completeness, freeman 1978.
Thus even if the country is to be divided into only 2 districts we have
NP-hardness.
It is conceivable this could be escaped
so each vote consists of BOTH a rank-ordering, AND a
set of approved candidates (is there any requirement that
these two be compatible)? Is the rank-ordering permitted
to include equalities or be a partial order?
wds
Election-methods mailing list - see http://electorama.com/em for list info
OK, Now that I finally understand the DMC voting system (which is quite
interesting),
I have a few comments and questions...
(15+3 reasons to love DMC from J.Heitzig F.Simmons)
1. Allows to distinguish important from minor preferences.
--range does too, only better.
2. Immunity from
Here is another question - will DMC lead to 2-party domination, or
not? To really answer this, it would help to understand optimal
voting strategy in DMC, which is probably beyond reach.
However, you may be able to just think about 3-candidate DMC elections
and thereby answer the question with
I suspect DMC will lead to 2-party domination.
See
http://math.temple.edu/~wds/crv/IncentToExagg.html
and consider this same example under DMC.
If the top-1, top-2, none, or all-3 of
the candidates are approved by all (I do not care which,
so long as you are consistent about it)
then B wins.
2. Immunity from second place complaints. Unlike in MinMax and
Beatpath, the DMC winner always defeats the candidate which would win
if the winner were not present.
--nice. Also true of range.
I'm sorry you're wrong here: If ballots are
voters A points B points
60 6040
6. Robustness against noise candidates.. cloneproof...
--also true of range.
Could you say more precisely what you mean here?
--Range voting is immune to clones in the sense that any number of cloned
candidates, all of whom get the same rangevote scores, can be added to the
scene,
and the
Do you suggest the election system should rather declare one of the
candidates which are not approved by anyone the winner than to demand a
new election because of lack of approved candidates. (I certainly don't
agree to that.)
--yes I do. The job of a single-winner election system is to
I regard A=B as a declaration of knowledge on the part of the voter.
he is saying I understand A and B and I think they have the same utility.
I regard A?B as a declaration of ignorance from the voter. he is saying
I do not know whether AB or BA, I really am clueless on this matter.
In some
--Also, now that I understand DMC has two kinds of monotonicity property, I
must report my admiration.
Heitzig: Why? Many Approval/Condorcet-hybrids are monotonic in both senses.
--fine. However, I had not seen any such methods previously,
I am ignorant. You also mentioned your favorite
Incidentlally, since you claim because you cannot explain the precise meaning
of a range vote
of 64 versus 65, therefore range voting is somehow horribel and inexplicable...
and you like DMC... I ask explain to me the precise meaning of
`I approve of Bush.'
Pretty difficult, isn't it? And
I do not agree these two things are equivalent, although they are related.
If a method exhibits favorite betrayal but only rarely then third
parties might be able to flourish. For example Coombs' IRV-like
method exhibits favorite betrayal. Would it lead to 2-party domination?
Really 2-party
cloneproof[strat2-revote]
A voting method hereby is cloneproof[strat2-revote] if W'=W when:
1. We hold an election with strategic voters, electing winner W.
2. we add clones (perhaps multiple clones) of some subset of the candidates.
3. The voters re-vote in the new election, again acting
Re your Weinstein idea that you would vote for candidates above the median
with
approval voting, since you do not believe in utility, I ask you to consider
A. Josef Stalin
B. Adolf Hitler
C. Genghis Khan
D. Jacques Chirac
where (say) ABCD in your opinion.
Depends on the
is actually quite pleasant. Try the demo at
http://math.temple.edu/~wds/crv/quickdemo/PresRadio.html
(and let me know if there are any browser problems. I have heard a rumor
there is a bug in early versions of netscape which may cause this
demo to be unattractive...)
wds
Election-methods
do you have a counterexample showing favorite betrayal for
Condorcet (winning vote, equality-rankings permitted, partial orders not
permitted)
?
(And if you do not, isn't that a good reason NOT to prefer DMC?)
wds
Election-methods mailing list - see http://electorama.com/em for
The original general-purpose 19-voter FBC example from the Center for Range
Voting web page
http://math.temple.edu/~wds/crv/IncentToExagg.html:
8:BCA
6:CAB
5:ABC
B wins under Condorcet Voting [Ranked Pairs variant, winning votes,
equality-ranking permitted] according to Eric Gorr's calculator at
seems to depend on having an exact tie. Therefore, this
counterexample is not as impressive as I thought.
wds
Election-methods mailing list - see http://electorama.com/em for list info
.
It then will be possible to proceed from there to have a genuine debate about
voting methods.
I am not going to debate voting methods with people who refuse to accept
probability theory,
believe that the sun revolves around the Earth, think Darwin is a phantasm, etc.
Warren Smith
Election-methods mailing list
robla: The problem with placing paramount importance on utility in voting
methods is not that it doesn't exist, it's that there's no systematic,
fair way of measuring utility.
--WDS: EXACTLY GOOD!!!
However, Heitzig has repeatedly and clearly stated that it does not exist.
I have repeatedly
robla: Warren, we don't agree. I said there is NO systematic, fair way of
measuring utility. I didn't say it's hard, I said it's impossible.
Ergo, for purposes of studying electoral systems, it might as well not
exist.
Using Bayesian regret on numeric utilities is begging the question. By
On 9/2/05, Andrew Myers [EMAIL PROTECTED] wrote:
I thought the folks on this list would find it interesting to see
some actual empirical data on how often cycles happen. I have data on
99 CIVS elections that have been run in which more than 10 voters
participated (max was 1749) and in which
To recount some recent history. I at first had this idea that all
Condorcet methods would lead to 2-party domination.
I in fact produced a proof of that (well, a proof of a related statement,
anyhow)
and put it on the CRV web site. Then one of the attacks on my proof (by Adam
Tarr) was that
my
is ICA the same thing as Smith//Approval? And if not, what is the matter with
ICA?
Election-methods mailing list - see http://electorama.com/em for list info
An analysis of the constitutional question is already available within the CRV
web site
http://math.temple.edu/~wds/crv/ConstVt.html
Also, re Robla's ludicrous range killing example illustrating range's
glaring defect, let me say this. You are perfectly free in the range
system to cast a
Yves in reply to wds's criticism of Robla's range killing example:
1- Sincerity doesn't exist in politic. As the vote itself, everything is
always strategic. The concept of democracy is to give the same chances to
all individuals to influence a collective decision.
--wds response:
first of all,
robla:
Incidentally, Range voting wouldn't have prevented slavery. Black
suffrage was a pretty important prerequisite which didn't exist back
then. Also, I don't think that a bunch of people who were willing to
secede from the union and fight a war on their own soil would express a
mild
wds:
Robla failed to mention that range voting *does* obey a weakened form of
the majority-winner criterion (call it WMW). Specifically:
If a strict majority of the voters regard X as their unique favorite,
then
they, acting alone without regard to what the other voters do, can force
MDDA *if* all votes are full rank orderings, is just the Smith set
and often yields a tied election. In fact often the Smith set is the entire
set of candidates. (In most of Australia full rank orderings - i.e. none
omitted -
are demanded by law.)
This seems a severe problem with MDDA and
Incidentally, Deluxe MDDA is probably even worse than un-deluxe ranked-ballot
methods
with respect to add-top failure, no-show paradoxes, and the like, because you
can
use the approval counts quite easily to set up bad scenarios where the new
voter creates (unfortunately for him) a Condorcet
MDDA fails add top. That is, if you add some identical honest votes ranking
A top,
that can harm A (e.g. by creating a Condorcet winner [who is not A]
who then wins, whereas previously there was a Condorcet cycle and A was the
winner
on approval counts).
Now this may not technically count as
lomax: Reading the scenario, I'm struck by how utterly unlikely it seems
that voters would actually behave in the way described. Essentially,
if practically the entire electorate decides to go on an exaggeration
binge, a bad outcome could occur. Well, duh!
On what basis is the claim made that
example of situation in IRV where truncating a ballot
is strategically desirable:
If your favorite is F but F is eliminated in round 1,
and the rest of your ballot is a no show paradox example
in which you are better off not showing up to vote,
then truncating your ballot
F --the rest
I just finished a paper on this topic. It is available at
#91 (at the end) of http://math.temple.edu/~wds/homepage/works.html
It incorporates quite a lot from EM-list members FW Simmons
and D Gamble, however they are not listed as coauthors since they declined,
or at least did not accept. The
yes, I'd be interesed in yor IR presidential election data etc.
election-methods mailing list - see http://electorama.com/em for list info
Arguably STV multiwinner electiosn are still of interest for single-winner
purposes since the FIRST winner is a single-winner IRV winner.
Gilmour is correct (I am happy to now learn) that Ireland is now posting
full vote lists in some (all?) STV elections on the www. I grabbed the
Dublin country
sorry, you are right that the first STV winner need not be the same as the IRV
winner.
Re NP-hardness polytime, all such results take
place in an asymptotic limit as some parameter or parameters describing
nput size tend to infinity. Say the input size is N bits; then
an algorithm is polytime
I have written up decent proofs of a lot of theorems about range voting
http://math.temple.edu/~wds/crv/RVstrat.pdf
http://math.temple.edu/~wds/crv/RVstrat.ps
this also includes the first explanation of optimum strategy in a large variety
of voting systems, includign COAF systems, range,
The MDD process is inherently incompatible with clone
independence. No matter what you do after MDD, it's possible
that cloning the winner, such that there's a majority-strength
cycle among the clones, will cause the win to necessarily move
to some other candidate.
--wait a minute. MDD
http://math.temple.edu/~wds/crv/FBCsurvey.html
I have added some stuff to incorporate Venzke's comments.
I was going to add a bit on Simmons' latest FBC-obeying lottery method.
But I think he is wrong so I deleted that bit.
In Simmons's method, you vote with a rank-ordering approval cutoff.
I have added this comment to the FBC-survey to address Venzke's correction
re clones:
But MDDA seems superior to MDDB because of its immunity to clones. That was
assuming all X's clones are ranked co-equal to X on all ballots. But if voters
can have slight preferences among the clones, then
Hi Rob.
Your little essay about how political parties form aka movie night
started out nice but got lame at the end.
(You also exhibit some high class knowledge of how to create web pages...
my web pages use old technology and I think simply cannot do the stuff you
did...)
Anyhow. To answer
I have examined this issue before in an unpublished paper whch I can
tell you about in separate email.
Anyhow, the thing is that some, but not other, Condorcet matrices are
actually achieveable as arising from actual sets of ballots.
Which ones are achievable? Well, you can tell by solving an
And I think you can construct a counterexample of this form:
for some fairly large number C of candidates, but only 2 voters,
make a Condorcet CxC matrix out of 2 random votes.
Can anybody conform or deny this?
wds
election-methods mailing list - see http://electorama.com/em for list info
no, I meant Bishop's alg would fail to fid ANY deconstruction even when one
existed. The problem was not, as Kislanko worries, that there might be
a non-unique solution. That is a valid worry, but I do not care about that
worry.
Also, to reply some more to Kislanko, he argued that a condorcet
Probabilities of favorite-betrayal lesser-evil (FBLE) situations in
3-candidate
ranked-ballot elections
Definition of FBLE situation: Call the election-winner A. An FBLE situation
then
occurs when some CBA voters, by switching to BCA (betraying their
favorite C)
can make B win (an outcome
Gilmour:
What I had in mind was if I vote 1, 2, 3, 4 (1 = most preferred,
the one I want to see win) for candidates A, B, C, D,
and you vote 100, 99, 2, 1 (1 = most preferred) for the
same four candidates, it would be fundamentally undemocratic if
your vote counted for more in determining the
Reply to Venzke Gilmour about Social Utility
Gilmour Venzke have again expressed the opinion that Social Utility is merely
yet another voting system criterion, on a par with
Monotonicity, Favorite Betrayal, Condorcet, etc, and therefore my preference for
it is unwarranted, mysterious, biased,
Venzke:
THIS IS useful TO know, since it means that the range voter's greater
ability to express himself (relative to the approval voter) isn't part
of why WDS feels social utility IS a useful measure. That means there's
no reason to discuss only range, since approval should justify the use
of
More results about computer simulations of elections - without need of a
computer!
--Warren D Smith--Dec
2005
We continue to consider the following cheapo model of simulating an election.
Each candidate to
each voter has a utility which
WDS: Thm 1.
Suppose all the voters magically know the identity X of the
max-summed-utility candidate.
Suppose each voter votes approval-style by approving of all candidates
with more utility
than f*U_X, where U_X is X's utility (to that voter) and f is a constant
(for example f=10% or
WDS:
Proof sketch:
Because that winner X will be Condorcet winner.
For each Y not in {A,B}, X is ranked above Y one-half of the time,
Venzke:
One-half isn't a reasonable guess for this. You're not considering the
equal ranking at all.
--WDS: wrong.
one-half was not a guess. It was an
Rob Brown:
How much do you want your vote to count (check one):
( ) As much as possible
( ) 90% of as much as possible
( ) 80% of as much as possible
( ) 70% of as much as possible
( ) 60% of as much as possible
( ) 50% of as much as possible
( ) 40% of as much as possible
( ) 30% of as much as
the exponential generating function is
exp(-x) / (2 - exp(x))
and the asymptotics for large N are C*N!/(ln2)^N for some constant
C = 0.36 very approximately.
I have not worked out the exact value of C and this estimate could
be off by a factor of 2 or so.
wds
election-methods
The Center for Range Voting is now located at
http://www.rangevoting.org
please alter ay hyperlinks you may have. Last I checked
google was still unaware of the new location, although the
old location does transfer you to the new one.
Warren D Smith
://www.rangevoting.org .
--Warren Smith
election-methods mailing list - see http://electorama.com/em for list info
OK, the hopefully-final version of my
Left-middle-right 3-candidate scenario
paper, which attempts an exhaustive computer simulation of
essentially all such scenarios, is now available at
http://math.temple.edu/~wds/homepage/works.html #95.
It concludes that range voting is clearly the
of the failure of IRV. Pro-IRV gorups have falsely
claimed there are no real world xamples of IRV failure.
I recommend joinign and endorsing CRV as usual...
Warren Smith
http://www.RangeVoting.org
PS. If you find more information about Mexico 06 that either supprots
or argues against (or clarifies
I entered the Debian leader election of 2006 into Rob Lanphier's
electowidget system and let 'er rip, and the result is
http://wikitest.electorama.com/wiki/Election:2006_DPL
and every voting method agrees Towns wins except range and approval both say
McIntyre wins.
I suspect RobL's method for
of elections
are nearly random.
4. plurality really fails a lot, numerous examples tabulated.
Warren Smith
http://www.rangevoting.org
election-methods mailing list - see http://electorama.com/em for list info
My article on wikipedia by that title has now been deleted.
I still do not understand what the complaints were about the article.
Supposedly it had a point of view although nobody was able to figure
out what it was, and/or it was original research although
nobody was able to figure out what that
thanks Abd. The google cached article is not the most recent version
which e.g included an additional section and considerbly larger bibliography,
unfortunately.
Also, Opabinia Regalis
intiialy gave a 1-line deletion request for my article and
gave no reasons whatever, however, after I later
To learn more about this fascinating story, see
http://www.rangevoting.org/ApisMellifera.html
It took me a while to understand the bees' method. At first I thought
it was kind of a hybrid of range voting, plurality voting, and negotiation.
But eventually it dawned on me, after reading
conversion method, subjectively speaking) and see
what happens.
Note: other endorsers and endorsements for range voting are sought.
Please contribute yours and/or recruit more...
Warren Smith
http://RangeVoting.org
election-methods mailing list - see http://electorama.com/em for list info
http://rangevoting.org/CondAppConflict.html
The Condorcet and Approval winners cannot differ if there is no tie
and if the Approval voters all place their approval thresholds strategically
under the guess it is going to be A vs C where A and C are the putative Approval
and Condorcet winners.
In
I attempted to write a careful discussion of these matters
for the general public in the CRV's usual inimitable style.
Anyhow, the interesting thing is that contrary to various
previous not-careful claims, it appears that the DH3 pathology
clobbers ALL condorcet methods, whether equalities are
Yes, it does bear some resemblance to A.Lomax's proxy ideas.
I too have devoted some thought to these ideas.
However, I suspect Lomax's ideas are better and Lanphier's worse. Or I
understand neither.
Specifically, as far as I understand it, with Lomax's proxies, you can
select anybody on the
In your example of 3 main rival candidates (A, B, C) and one dark horse
candidate (D), you said that range voting prevented the dark horse from
winning. Graphically speaking, there would be a triangle formed by the
three main candidates , while the dark horse would lie somewhere outside
of it.
My
I have made a new CRV web page devoted to answering the questions
http://RangeVoting.org/WVmore.html
Warren Smith
http://RangeVoting.org
election-methods mailing list - see http://electorama.com/em for list info
I'm not sure your IRV restricted to Smith set BUT do NOT eliminate the
non-Smith candidates, is better than the pre-elimination.
Think you could make arguments either direction.
But I guess this is an argument for BTR-IRV; with BTR-IRV you do
not have to worry about that issue, since
Sorry, my last email was in error: BTR-IRV can entirely eliminate the Smith set
and elect some nonmember.
wds
election-methods mailing list - see http://electorama.com/em for list info
1 - 100 of 196 matches
Mail list logo