Talk:Strategic voting

Congratulations to all involved - I think I started/contributed to this page (or one like it) some time ago: and now it is very much better than anything I could possibly have done.

Chris Baker

---

It can even be encouraged by political partys: If a party will not be strong enough by itself to form a government, but could form a coalition government if another party would pass the 5-percent-clause it will probably encourage tactical voting. (As did happen in Germany in 1984)

Okay, I understand how Germany's system works now, and see how this is tactical voting. I'll try to work this in on one of the PR pages. --DanKeshet

---

The section of 'Elite manipulation' is inappropriately titled and, to me at least, seems to be beside the point somewhat. Yes, people can be encouraged to vote tactically, but you seem to be making unnecessary value judgements here.

I didn't mean to make a value judgment there; I was just copying from the academic literature. (As a frequent campaign activist, I find it amusing that anybody would call me "elite".) I've changed it to something I hope will be more palatable.

Secondly, the article is a little too technical, too quickly. Without some mathematical background, I think you'd struggle to get much useful out of it. --Robert Merkel

I'll try to factor out the game-theory discussion into appropriate pages (like the G-S theorem page) and leave this page as an overview of how tactical voting occurs in real-life elections.

Incidentally, I've just started reading the academic literature on voting systems, and I'm now embarrassed at all the pages I've put up from the activist perspective. --DanKeshet

2-1 Voting

It seems this important voting strategy, which merits inclusion in Wikipedia, doesn't fit in either strategic nomination, vote swapping, or electoral fusion:

2-for-1 Voting

by Bruce Ackerman

In November, Americans won't be casting their ballots directly for George Bush, John Kerry or Ralph Nader. From a constitutional point of view, they will be voting for competing slates of electors nominated in each state by the contenders. Legally speaking, the decisions made by these 538 members of the Electoral College determine the next president.

In the case of Mr. Bush and Mr. Kerry, electors will be named by each state's political parties. But Ralph Nader is running as an independent. When he petitions to get on the ballot in each state, he must name his own slate of electors. While he is free to nominate a distinctive slate of names, he can also propose the very same names that appear on the Kerry slate.

If he does, he will provide voters with a new degree of freedom. On Election Day, they will see a line on the ballot designating Ralph Nader's electors. But if voters choose the Nader line, they won't be wasting their ballot on a candidate with little chance of winning. Since Mr. Nader's slate would be the same as Mr. Kerry's, his voters would be providing additional support for the electors selected by the Democrats. If the Nader-Kerry total is a majority in any state, the victorious electors would be free to vote for Mr. Kerry.

This plan is consistent with the original understanding of the founders. When they created the Electoral College, they did not anticipate the rise of the party system; they expected voters to select community leaders who would make their own judgments when casting their ballots for the presidency. In designating Kerry electors rather than insisting on his own slate, Mr. Nader would be giving new meaning to this tradition that refused to view electors as simply vehicles of a candidate's will. In effect, he would be enabling his supporters to rank their choices: Mr. Nader first, Mr. Kerry second. (full article)

Should a new article be created? If so, how should it be named? Sir Paul 03:21, Jun 30, 2004 (UTC)

I would think this could go into the U.S. Electoral College article. There is a detailed discussion there now which this would fit well into. A brief summation of the general principle involved would also improve the broader Electoral College article. - toh 08:09, 2004 Nov 3 (UTC)

---

In the US the people who tried to fight the injustices of the electoral college and did vote swapping sites in the 2000 election banded together and now have a site, http://www.votepair.org

STV Tactical Voting

There is significant content in the Single Transferable Vote article about tactical voting in STV systems. Should it be moved here, or linked from here, or both? Scott Ritchie 20:43, 4 May 2005 (UTC)

Arrow's theorem

"Since Arrow's impossibility theorem proves that any voting system is arguably undemocratic in at least some case"

Could anyone please explain to me why this claim deserves to be here? I would take it off. Arrow and Gibbard-Satterthwaite show that any 'democratic' method is prone to manipulatable situations. Not that voting is un-democratic. It says life is complicated, not that life is wrong. mousomer 09:25, 24 October 2005 (UTC)

Borda vulnerability

"The Borda count has a both a strong compromising incentive and a large vulnerability to burying." Is this really true? If I'm not mistaken, then a large proportion of voters need to employ both in order to make a significant change - see Borda count#Potential for tactical voting. Common Man 11:13, 24 January 2006 (UTC)

Well, it's easier than you think. The crux of the matter is this: if you have one very favourable candidate (who's position is very strong), there is very little that tactical voting could do. The trouble is that the chances for that are not great. In most real life situations, you have several candidates, none of whom has a definite magority behind him. If you do the statistics, you could see that the most likely situations are where there is a tie or near tie between the candidates. It is exactly under these near-tie situations that tactical voting becomes relevant. mousomer 12:14, 28 January 2006 (UTC)

changed "vote swapping" in see also to "vote pairing" since that's where the vote swapping article has moved. Snowden666 18:18, 23 March 2006 (UTC)

explaining edit

I changed the line "that is, might not select the same outcome every time it is applied to the same set of ballots" to "that is, might not select the same outcome every time it is applied to the same set of voter prefrences" Because most methods will choose the same canidate If the ballots are the same, but with a diffrent amount canidates could choose a diffrent one.24.237.198.91 01:29, 23 November 2006 (UTC)

Brams and Herschbach

I see a reference which is quite non-standard (to say the least):

Brams, Herschbach, "The Mathematics of Elections" (sic?), Science (2000)

Apparently it was added a long time ago and never corrected or verified. I guess it refers to: "The Science of Elections" available online here:

http://www.sciencemag.org/cgi/content/summary/292/5521/1449

This reference is correctly listed in the following article of Wikipedia: http://en.wikipedia.org/wiki/Approval_voting

Best regards

Fioravante Patrone http://www.diptem.unige.it/patrone/default.htm

How to make manipulation computationally hard?

I think the following or similar reference should be added:

Universal Voting Protocol Tweaks to Make Manipulation Hard (2003), Vincent Conitzer, Tuomas Sandholm

http://citeseer.ist.psu.edu/conitzer03universal.html

http://www.cs.cmu.edu/~sandholm/voting_tweaks.ijcai03.pdf

From the "Abstract"

Voting is a general method for preference aggregation in multiagent settings, but seminal results have shown that all (nondictatorial) voting protocols are manipulable. One could try to avoid manipulation by using voting protocols where determining a beneficial manipulation is hard computationally. A number of recent papers study the complexity of manipulating existing protocols. This paper is the first work to take the next step of designing new protocols that are especially hard to manipulate. Rather than designing these new protocols from scratch, we instead show how to tweak existing protocols to make manipulation hard...

The protocols become NP-hard, #P-hard, or PSPACE-hard to manipulate...

These are the first results in voting settings where manipulation is in a higher complexity class than NP (presuming PSPACE ≠ NP).

I need a feedback about this. Or, if sure, please add such or similar reference...

--Y2y 11:06, 5 March 2007 (UTC)

'Tactical' v 'Strategic'

It seems to me worthwhile to point out -- and I thought I once had, but perhaps not here -- that while 'strategic' and 'tactical' are usually used to distinguish behaviors in a disjoint set, here they counterintuitively mean the same thing, which is the topic of the article. If in fact everyone agrees that a) this is true and b) it's worth mentioning... where should it go? I'd think a parenthetical in the lede, but the lede diverges pretty rapidly right now from graf to graf...
--Baylink 19:23, 5 May 2007 (UTC)

I think it's worth mentioning too, tactical and strategic means two differents things, and the article is named "Tactical voting" after all, explaining "tactical vs strategic" is necessary. However as english is not my native language and I'm novice in voting studies, I don't know where it should go too. --88.177.158.231 (talk) 02:41, 8 December 2010 (UTC)

Rewrite

I've put a rewrite tag on this article, because the overall tone appears to be somewhat negative. I've attempted to clean up the introduction, however the rest of the article needs work. Addhoc 16:49, 7 August 2007 (UTC)

Can you cite specific examples? It looks great to me. ←Ben^B4 16:36, 13 September 2007 (UTC)

Removal of Tactical Voting argument re Approval Voting.

Because Approval Voting can be understood as asking the voter to set an Approval cutoff, and the votes then express the "sincere" opinion of the voter with regard to that cutoff, the application of "tactical" voting as meaning an insincere vote for desired effect is problematic. There is no advantage to insincere voting in Approval.

Voting "strategy" is another matter. The thinking that there is tactical voting in Approval is based on an idea that there is some absolute Approval or Disapproval for candidates, whereas, in fact, approval is always a relative term, dependent upon the alternatives presented. So in a real election, the voter will make judgments as to where to set the Approval cutoff, dependent upon what the voter considers possible. A great deal of confusion is caused by the name "Approval Voting," and this was shown in what I removed. The voter may set the approval cutoff anywhere without being insincere. (Or, more accurately, if the voter votes insincerely, by reversing preferences, the voter harms his or her own expected outcome.) Abd 03:10, 7 October 2007 (UTC)

Contradiction with Approval voting

Tactical voting suggests that Approval voting is (nearly) immune to strategy, but approval voting says exactly the opposite. Both claims have references. Is there an expert in the house? Vectro 20:18, 8 October 2007 (UTC)

Approval voting is immune to deceptive claims (if a voter lies about their actual preferences then they hurt their chances of someone they prefer getting elected in all cases). However, Approval voting is VERY strategy dependent. For instance, if there are 3 candidates (A, B, C) and

40% of voters prefer A > B > C

29% of voters prefer B > C > A

31% of voters prefer C > B > A

Each voter needs to decide if they want to vote for their best candidate only, or vote for their top two candidates. A's supporters probably realize they are in the lead, and so will vote only for A. This puts people voting for C and B in a bind- as long as most of them vote for B and C, they will beat A. Most people voting for B or C will vote both, just to keep A out. The temptation exists, however, for just a few to vote for C only or B only- if their tactic succeeds, B will triumph over C (or vice versa). If too many succumb to the temptation, though, A will win despite being the worst choice for the majority of the people.

Thus, if I support B, I'm not sure whether voting (C and B) or just voting (B) will help me- I have two strategies to choose from, and have to pick one of them. I do know, however, that it's always in my interest to vote for B, and it's never in my interest to vote for A. (In a Plurality voting system, It may be in my best interest to vote for C instead of B, despite my preference for B).

Hopefully that explains why approval voting requires strategy, but never deception. Paladinwannabe2 19:38, 11 October 2007 (UTC)

Since we are defining tactical voting as 'voting insincerely' instead of 'voting strategically' Approval voting does not conflict with tactical voting. I will remove the contradiction flags. If we want to change the definition of tactical voting, then tactical voting could apply to approval voting. Paladinwannabe2 20:20, 11 October 2007 (UTC)

Definition of "Tactical voting" and application to Approval and other cardinal methods

From the introduction: "In voting systems, tactical voting (or strategic voting or sophisticated voting) occurs when a voter supports a candidate other than his or her sincere preference in order to prevent an undesirable outcome."

Now, how do we apply this definition to Approval? If we take the strict definition, without considering context, we would come up with the conclusion that if a voter adds a vote for any candidate other than the favorite, this is "tactical voting." However, that additional vote might be "sincere," and "sincere voting" is generally considered to be the opposite of "tactical voting." The definition is flawed, and the reason is that it was written for ranked methods, where "tactical voting" implies preference *reversal*. Warren Smith explores this issue on rangevoting.org, developing language to address a difference between preference reversal, which is fully insincere, preference expression, which is fully sincere, preference equality if the voter actually considers the candidates equal -- or possibly the preference is below some level of significance -- as also fully sincere, and preference submersion (my language), where the voter has a significant preference, would clearly choose one candidate over another in a direct contest, but votes for both for strategic reasons. I.e., the favorite doesn't have a snowball's chance in hell of getting elected!

Applying the definition as it stands to Approval takes us into some weird places. Suppose there is an election between A and B, and the voter intends to vote for A. Oops! Last minute, C announces a write-in campaign. If the voter votes for A but would actually prefer C, is this a "tactical vote"? Make this into an Approval election, and most voters that prefer A to B, and think C is unlikely to win, would probably vote for A and C, and consider this a sincere vote, even if the preference for C is strong. They will look at the election in terms of what is realistically possible, and make a choice among the realistic choices. If, in fact, both A and C gain a majority, they might regret their vote for A. This contingency, though, by the conditions of the problem, is rare.

Assuming that equal ranking in the presence of a preference is "tactical voting," then, leads to problems. It defines what the voter may consider a perfectly sincere vote, thinking of Approval as a matter of setting an approval cutoff, and the voter sincerely set that cutoff, as a "tactical vote," and this is then used in arguments to claim that Approval encourages "insincere voting." Which is, ahem, bovine refuse.

In the contrary direction, some of the same people will look at Approval and consider it vulnerable to tactical manipulation by small groups of voters, based on the idea that these voters actually approve of A and B, but vote only for their favorite, A, and then that favorite wins, whereas if these voters had voted "sincerely," B would have won.

Yet, in this case, this block of voters simply voted for their favorite. How is this "tactical voting"? It seems that there are people, even academics, playing with the definition to suit their theses.

When Brams claims that Approval is not vulnerable to tactical voting, he is claiming that *preference reversal* is never rewarded. Whereas, when the concept of tactical voting is applied to ranked systems, preference reversal is always the case.

The definition in the article *can* be read as requiring preference reversal. We simply read "supports a candidate other than his sincere preference" to mean "raises in rank or rating above his sincere preference." However, clearly, the definition wasn't written to contemplate equal ranking. Same problem with the Majority criterion.

I intend to do some more research and edit the article accordingly, comment invited, and, especially, any reliable sources on the definition of tactical voting that aren't already in the article. --Abd (talk) 17:34, 19 January 2008 (UTC)

I agree that the lead description of tactical/strategic voting is poorly written. The logical application of that description to various voting systems is unclear. There is room for improvement. However, it may be unrealistic to expect to find a common definition or a universally applicable definition in the literature.

There are at least two core issues here: what is tactical/strategic voting and what is sincere voting. Some definitions of the former are dependent on the latter. In any case, addressing the issue of sincere voting in the general case is needed on Wikipedia and this article is an appropriate starting point. The approval voting and range voting articles deserve a discussion of sincere voting specific to those voting systems.

(It is important to not transform Brams' claims about tactical/strategic approval voting by dropping the restrictions and qualifications that he attached to those claims.)

DCary (talk) 20:24, 19 January 2008 (UTC)

That's correct. However, the article places in opposition Brams' claims and those of the other authors (they "disagree"). Are they really in opposition, or are they using different definitions and conditions? Further, there is a problem here. The citation to the article in Public Choice is to an abstract that does not state what the WP article states. Presumably it's taken from the text of the Public Choice article. But that isn't easily accessible. Presumably, the person who inserted the material here has access to the actual article; therefore I'm asking for quotation of sufficient text to establish the context and meaning of the Public Choice quotation.

There is another problem that the Brams article is not actually cited, though I may be able to find it (I think I may have seen it somewhere). The source should be properly cited in our article.

--Abd (talk) 21:27, 19 January 2008 (UTC)

Okay, it always pays to check. There was an editorial in Science by Brams and Herschbach on Approval Voting.[1] It does not make the claim in the text of this article. Don't you love arguments based on what one side didn't say? However, that editorial does note some well-known problems with ranked methods, but it doesn't make the comparative statement. I think what Brams has written elsewhere was the basis, so the article is incorrect and should definitely be fixed. I don't yet know what to fix it *to*, yet! --Abd (talk) 21:35, 19 January 2008 (UTC)

Ooops! That was a summary, it was so long I thought it was the editoral. At least that's how it looks right now. Sometimes Brams has copies of articles elsewhere, I might be able to find it. --Abd (talk) 21:40, 19 January 2008 (UTC)

Expressive Voting

There has been a considerable academic literature develop in the past decade on "Expressive Voting" -- but I don't find it covered in Wikipedia at all. It might, at least, merit a section in this article on Strategic/Tactical Voting. I did a quick search on Google Scholar and found this paper, with a serviceable definition, at the top of the list as "most cited" (it indicated there had been 68 citations of this particular article in the scholarly literature). From the abstract:

"There are two rival accounts of rational voting ...: the mainstream instrumental account, that sees the vote as a revelation of preference over possible electoral outcomes, essentially analogous to a market choice; and the expressive account, that sees the vote as expressing support for one or other electoral options, rather like cheering at a football match. This paper attempts to lay out some of the implications of the expressive account ..." Brennan, G (1998). "Expressive voting and electoral equilibrium". Public Choice. 95 (1). Kluwer: 149–175. ISSN 0048-5829. {{cite journal}}: |access-date= requires |url= (help); Cite has empty unknown parameters: |month=, |quotes=, |laydate=, and |laysource= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help); Unknown parameter |laysummary= ignored (help)

I think this idea is a bit different than what is reflected in the article today, yet may be considered a form of tactical voting. What do others think? N2e (talk) 16:35, 12 February 2008 (UTC)

What you describe is interesting and related to voting, but it does not seem a part of tactical voting. See the first sentence of the article: "In voting systems, tactical voting (or strategic voting or sophisticated voting) occurs when a voter supports a candidate other than his or her sincere preference in order to prevent an undesirable outcome."

It seems that expressive voting would be where a voter votes for a candidate they really would prefer, but perhaps has no chance to actually prevail in an election. Maybe a new article is needed. —Preceding unsigned comment added by 71.163.232.181 (talk) 15:08, 1 March 2008 (UTC)

Rational Choice Voting is not the equivalent of Lesser-of-Two-Evils

In the section discussing the rational choice voting formula, the author said that the rational plurality strategy is to choose the best of the "frontrunners". In fact, the highest prospective rating can theoretically belong to any candidate, including those who could under no definition be called frontrunners. The misconception was called, correctly, the "common" strategy. I kept that observation but corrected the context.Tomblikebomb (talk) 20:19, 8 October 2008 (UTC)

What??

These two statements from the article are directly contradictory:

"As the United States has, by and large, a two-party system, there is almost no opportunity for tactical voting to occur in general elections."

But earlier in the article:

"Compromising (sometimes "useful vote") is a type of tactical voting in which a voter insincerely ranks an alternative higher in the hope of getting it elected....Duverger's law suggests that, for this reason, first-past-the-post election systems will lead to two-party systems in most cases."

Someone who has more experience with this needs to either rewrite one of these sections, or explain in the article how this isn't a contradiction. Disputulo (talk) 03:57, 5 November 2008 (UTC)

Actually, there are experts already looking at this article. Problem is, a lot of expertise exists that isn't verifiable, at least not easily, from the published literature. In researching the question of whether or not Approval voting and Range voting are "vulnerable" to "tactical voting," I ran into a problem. The original literature on strategic or tactical voting assumed preference reversal. I.e., "supporting" a candidate who is not the favorite meant "favoring" that candidate over the favorite, i.e., by ranking the less-liked candidate higher. Because voting systems theorists were largely fixated on preference voting, using a ranked ballot, it was simple and easy to define "tactical voting."

Disputulo is correct. Plurality in a two-party environment, requires tactical voting for anyone who would prefer someone other than a frontrunner, with a penalty of the vote being mooted if the voter votes sincerely. Considering that ballots where write-ins are allowed have an unlimited number of virtual candidates, whom the voter could, in fact, vote for, the norm is tactical voting. That is, nearly everyone is compromising.

However, Approval, and then Range, introduced a new possibility that wasn't much considered previously: equal ranking. If I prefer A, but I'm willing to accept B, and I vote for A and B, equally, am I voting tactically? By the classical definitions, which required preference reversal, no. And this would mean that Approval Voting wasn't "vulnerable to tactical voting," i.e., voting with reversed preferences, which would mean voting for B only when I prefer A, provides no advantage to the voter. If the voter prefers A, to B, and the voter has decided to vote for B (perhaps B is a frontrunner, A is not), then it is always an advantage, or at least harmless, to vote as well for the favorite, A.

But then new definitions of strategic voting were invented that would cause Approval to appear vulnerable. Quite simply, these definitions mean that if you prefer A to B, you *must* rank A over B, no matter how small the preference is, or you are "voting tactically." Essentially, preference-oriented theorists crammed Approval and Range into preference-thinking boxes, and did not consider the extended implications.

Much of the problem comes from considering there to be some absolute state in the mind of the voter with respect to each candidate, i.e., "Approval." The name of the method encourages this error. Approval is relative, not an intrinsic characteristic of a particular candidate for a particular voter; whether or not we "approve" of A will depend on what other options exist, and how likely we consider them to be elected. If "tactical voting" means "voting with a knowledge of what is practically possible," yes, sure, all known voting methods reward tactical voting. But that's not a bad thing! "Vulnerable to tactical voting" is used as a perjorative. Of course, that's not how "tactical voting" is defined in this article. The definition as it is:

In voting systems, tactical voting (or strategic voting or sophisticated voting) occurs when a voter supports a candidate other than his or her sincere preference in order to prevent an undesirable outcome.

There may have been some degree of conflict over this. The way it is worded, voting for B if you prefer A -- no matter how slight the preference -- could be defined as "tactical." However, with ranked systems, "support other than" has a clear meaning: rank higher than. Only when equal ranking is allowed does the meaning become possibly unclear. When Woodall defined his Later-no-harm criterion, he noted that a referee for the article had indicated severe displeasure about this being considered a positive criterion. Why? Well, the political process is largely a process of finding broadly acceptable compromises, and a method concealing possible compromises interrupts and inhibits that process. Approval voting allows voters to indicate sets of acceptable outcomes. Range voting allows them to do so with more precision (Range voting is really Borda count with a set number of ranks equal to the number of candidates or greater, and equal ranking allowed). Again, in considering Range to be vulnerable to tactical voting, it is assumed that there is some "sincere rating" for candidates, and the voter is voting tactically if, to improve the outcome, the voter does not vote precisely that. However, in fact, these are simply votes. They are actions, not sentiments.

The voter looks at the realistic possibilities, and places, if "sophisticated," the voting power where it is most likely to improve the outcome. It's really quite simple, and most people already know how to do this. Over Plurality, Range and Approval allow the voter to *improve* the sincerity of their vote without damage, and hybrid methods allow *fully sincere* votes without harm. The standard "strategy" or "tactic" is simple in elections with two frontrunners, which is the norm in public elections, true three-way races being quite rare. Vote maximum rating for your favorite frontrunner, vote minimum rating for the worst frontrunner, then vote all the other candidates wherever you please, never reversing preference. Or vote almost max or almost min for the frontrunners, if you want to express a preference outside that range for, say, your true favorite. (Hybrid methods would allow you to express preference without diluting vote strength.)

Now, try to find reliable source for this. Good luck. In recent years, there has been an explosion of understanding and knowledge regarding voting systems theory, and almost all of it has taken place on the internet, mostly in mailing lists. What does Mathematics and Democracy (Brams, 2008) cite? Not uncommonly, mailing list posts by, say, Ossipoff and Smith. Likewise Gaming the Vote (Poundstone, 2008). Some of the older published, peer-reviewed literature is just plain wrong. Or, more often, effective misleading, i.e., stated within one context where it would now be understood within another. For example, Arrow's theorem excludes, even from consideration, equal-ranking systems. Full preference ranking is assumed. --Abd (talk) 19:58, 5 November 2008 (UTC)

That is one heck of a dissertation covering a lot of information barely related to the issue. I'm glad for your passion. The first paragraph is by far the most useful. If everyone is in agreement that they conflict, which I think they do, then they need to be fixed, which is really as easy as removing the unverifiable and unecyclopedic first statement that the US races are immune. I would say that many US elections have at least some races involving more than two candidates, let alone primaries and caucuses. (plus, think of all the school board elections, most have at least one independent or third party in the mix) Disputulo (talk) 04:19, 6 November 2008 (UTC)

Passion, smasshion. I just happen to know a bit about the topic. And it's difficult to write for Wikipedia on it, because what is well-known isn't necessarily published. However, I removed the unsourced statement, after writing the tome above. Consider this: most U.S. election allow write-in candidates, and one may write in any eligible person, who does not usually have to be a declared candidate. If there is only one candidate, and one prefers someone not on the ballot, there is little or no harm in writing in a name. (There might be harm if there is another active write-in campaign.) But if there are two candidates on the ballot, then writing in the name of one's true favorite (not on the ballot) will quite likely waste the vote; only if there is an active write-in campaign is this not the case. I have seen no voting theorists consider the issue of write-ins. In fact, write-ins sometimes win elections, they are not insignificant. --Abd (talk) 00:12, 7 November 2008 (UTC)

IRV vulnerability

I don't see how instant runoff voting is vulnerable to the push-over strategy. I'm not so confident that I'll edit it out, but could someone doublecheck this and explain it here or in the article if I'm wrong? 66.41.154.228 (talk) 05:03, 16 December 2008 (UTC)

I've been researching these things for decades now, and I've never heard of a "push-over" strategy or tactic against runoff voting. In the past, we have had trouble with complaints being based on small sets of contrived, anecdotal examples. The peer-reviewed sources we have at Instant-runoff_voting#Tactical_voting seem entirely consistent with the likelihood that someone just made this up. GetLinkPrimitiveParams (talk) 04:12, 1 February 2009 (UTC)

"Unmitigated evil"?

I would like to take issue with the following citation:

Authors of a 2004 paper in Public Choice disagreed, saying, "AV is one of the most susceptible systems to manipulation by small groups of voters (for example, small, maverick groups could determine the AV outcome.)"

If you read the source paper, then read the paper it's responding to, then read the original paper that started the thread, you eventually realize that for Saari a voting system is not a way of choosing a winner, but a way of rank ordering all of the candidates. So in this sentence, Saari could be saying (it's really quite unclear) that Approval is very susceptible to manipulation that would change last and second-to-last place "results" without changing the winner of the election. Of course: if both got only their own vote, any voter could simply vote for one of them on a whim and break the tie. This statement says nothing about the kind of tactical voting that is the subject of this article.

If Saari himself comes on here and clarifies his statement, I might consider it a RS for this article. Until then, I'm going to be bold and remove the statement. Homunq (talk) 04:43, 15 June 2009 (UTC)

2003 California Recall

The paragraph highlighting the events leading to the 2003 California Gubernatorial recall election is wrong. California, at the time, did not have an open primary system for governor, where by Democratic voters could participate in the Republican primary between Richard Riordan and Bill Simon. Independent expenditure campaigns that were favorable to Davis' candidacy ran opposition against Riordan, so that Republican voters would select Simon as their gubernatorial candidate for the general election. —Preceding unsigned comment added by 69.235.223.217 (talk) 10:11, 7 May 2010 (UTC)

I fact tagged it, I'm pretty sure you're correct (At least today as I voted in the CA gubernatorial primaries it was closed - one can only vote for candidates in a single pre-registered party). According to [2] CA has had a "modified closed primary" since 2001 and thus would have been in effect in 2002 and 2003. This was changed from the blanket primaries that the US Supreme Court declared unconstitutional. I will do some more research and determine whether I should change or excise this information from the article Valley2city^‽ 01:42, 9 June 2010 (UTC)

Real life examples

The Winchester by-election is not a good one, so I have commented it out and will delete it unless there's a huge demand for it to stay. There were only around 6.5k Labour voters at the previous general election, so (most of) them switching cannot be the cause of a 21k majority! Less than 6k of the new LibDem majority can be attributed to such tactical voting and far more of it came from previous Tory voters saying, in effect, to Malone 'you lost, stop whining'.

If you want a UK by-election example, there are numerous better ones (eg Greenwich 1987, where on polling day the SDP/Liberal Alliance campaign called on voters who said they would vote Conservative to remind them to vote on the basis that their favoured candidate was known to be third in opinion polls, so many of them would vote for the SDP candidate who could beat Labour. The result was SDP/Liberal Alliance candidate up 28%, Labour candidate down 4%, Conservative down 24% = SDP/Liberal Alliance win... down to tactical voting.)

There are probably enough examples in the section already though. Lovingboth (talk) 10:10, 16 May 2010 (UTC)