EVERY deterministic voting method is susceptible to tactical voting. That is mathematically proven. With Approval Voting however, voters NEVER have to fear supporting their sincere favorite candidate. It passes the Favorite Betrayal Criterion unlike virtually every ranked system. If you think it's bad for voters to have to ponder whether or not to support a SECOND favorite candidate, imagine how bad it is to have a system in which you are generally encouraged not to support your FAVORITE unless that candidate has a strong chance of winning. See ScoreVoting.net/FBCsurvey.html

The general point here is, what about the case where there's more tactical voting with Approval Voting than with some other system. First of all, there's no evidence for that. For most ranked voting methods, voters generally use the Naive Exaggeration Strategy, meaning that tactical voting is used by something like 80-90% of voters. And that's actually, coincidentally, the generally correct strategy for ranked systems.

Here's a page in which we compared the performance of voting methods based on whether 100% or 50% of voters are sincere. Note that Approval Voting performs better with fully half of the voters being tactical than most other methods do with NO tactical voting whatsoever. This accounts for the differential tactical-ness you propose.

Lastly, there are two theorems that specifically describe how tactical voting with Approval Voting actually leads to mild (some would say, good) results.

Tends to elect Condorcet winners when they exist: ScoreVoting.net/AppCW.html

Maximizes the number of pleasantly surprised voters: ScoreVoting.net/PleasantSurprise.html

No, Range Voting is not "broken". Perhaps 90% of voters will vote "approval style" with it, but the 10% who get more satisfaction out of a more expressive honest vote will be happier, PLUS they will in the process cede some power to the 90% who were strategic, causing those 90% to ALSO have a better expected satisfaction. So the average voter, across the entire electorate, will be better off with Range Voting. Range Voting is also increasingly known as Score Voting.

ScoreVoting.net/ShExpRes.html ScoreVoting.net/StratHonMix.html

Score Voting is superior to all ranked voting methods, even ones which have never been invented.

ScoreVoting.net/BestVrange.html

Instant Runoff Voting is a very poor system, plagued by paradoxes and excessive complexity.

Approval Voting is better according to essentially every known metric.

www.electology.org/approval-voting-vs-irv

Borda is actually quite terrible. Extremely vulnerable to tactical voting. Borda himself said, "my system is meant for honest men".

http://scorevoting.net/BordaExec.html (Approval Voting is the simplest form of Score Voting (aka Range Voting), so those points apply to Approval.

IRV (what you use in the Australian House) is pretty lousy, and maintains two-party domination (save for the rare case where you recently got a single Green elected out of 564 seats).

ScoreVoting.net/AustralianPol.html

And that's in spite of having proportional representation in your Senate, in which Greens win a significant number of seats.

A comprehensive rundown of the MASSIVE superiority of Approval Voting is here:

www.electology.org/approval-voting-vs-irv

IRV is much worse in that regard. With Approval Voting, you would be strategically advised to vote for the Republican. But then you could still safely vote for the Tea Party candidate. And if it turns out enough people preferred the TP candidate, then he can win even if voters like you thought he had no chance.

But with IRV, your best tactic would be to insincerely rank the Republican in first place. And with a ranked ballot, most voters naively do that anyway, as we can tell from decades of use of IRV in e.g. Australia. That means if you don't think the TP candidate can win, it becomes a self-fulfilling prophecy. This is why Approval Voting is far far more fair to minor party and independent candidates than any ranked system.

www.electology.org/approval-voting-vs-irv

Approval Voting is superior to Instant Runoff Voting in just about every conceivable way. www.electology.org/approval-voting-vs-irv

Two reasons this doesn't hold.

First, the current system is susceptible to eliminating votes by spoiling them. Say you're a crooked poll worker in a largely black neighborhood in Florida. Just take some ballots for Gore and also mark Nader. Now they are spoiled, and must not be counted. With Approval Voting, that wouldn't be possible. You could also cast a vote for e.g. Bush in that case, but that would be a red flag compared to a ballot which marked e.g. Gore AND Nader.

Another major consideration is relative importance of these things. Approval Voting roughly doubles the quality of democracy, as measured via extensive Bayesian Regret calculations. On average, that's a much bigger benefit than typical fraud levels are a detriment.

Put another way, look at Gore/Nader/Bush in 2000. Even with all the fraud that happened, Gore was a mere 537 votes behind Bush in Florida. But 97,488 people voted for Nader. If even a tiny fraction of them would have also voted for Gore, given Approval Voting, then Gore would have won.

Bad voting methods have a statistically much greater effect on elections than fraud. Of course, if fraud is bad enough it doesn't matter WHAT voting method you use. But various factors tend to limit how extensive fraud can be.

Here is a page where Warren D. Smith, the Princeton math Ph.D. who authored most of the material at ScoreVoting.net, tries to get a ballpark estimate of the relative importance of various election reforms.

ScoreVoting.net/RelImport.html

Absolutely not. Say there are 3 candidates: X, Y, and Z, all tied.

You vote for X while I vote for Y and Z. Now they are all still tied. Our ballots had an equal but opposite effect there.

Mathematically speaking, every voter is approving or disapproving of every candidate. It's MORE equal than our current system in that sense.

More here: http://www.electology.org/hb-240#TOC-Doesn-t-Approval-Voting-violate-one

Proportional representation is only possible in multi-seat elections. When you're talking about single-seat elections (e.g. senator, mayor, president), Score Voting is essentially the best you can do. Approval Voting is the simplest form of Score Voting. Also, you have to be political realistic here. Even this mild change to the ballot rules, to enact Approval Voting, will be extremely difficult to pass. You want to sell a state largely ruled by Republicans on a system like Single Transferable Vote, with complex re-weighting rules? You need to start small and be pragmatic. That's what the bill's author is trying to do. Moreover, PR is extremely difficult to enact (nearly impossible) in a political climate dominated by two parties. Many feel that you won't get PR unless you first end the duopoly as a prerequisite. Ranked methods appear impotent in that regard, whereas Score/Approval Voting seem conducive to (eventually) allowing for more than two parties. So if you eventually want PR, getting Approval Voting ought to be your #1 priority. ScoreVoting.net/PropRep.html

This is simply false. Approval Voting is one of the best voting methods as measured by Bayesian Regret, and is extremely resistant to tactical voting. www.electology.org/tactical-voting www.electology.org/threshold By contrast, most ranked voting methods fall to the Naive Exaggeration Strategy. ScoreVoting.net/NESD.html This is especially problematic for Instant Runoff Voting. www.electology.org/irv-plurality

Score Voting (aka Range Voting) is certainly better than Condorcet. ScoreVoting.net/BayRegsFig.html ScoreVoting.net/CondBurial.html ScoreVoting.net/AppCW.html It is mathematically proven that the Condorcet winner is not necessarily the "right" winner (social utility maximizer), therefore the very POINT of Condorcet methods is the wrong one. They just happen to be pretty good. But since, in practice, most voters will naively exaggerate when they have a ranked ballot, essentially any ranked voting method will degenerate approximately into Plurality Voting in practice. ScoreVoting.net/NESD.html

Exactly. And it has lower Bayesian regret, which is more important. (The Condorcet winner is not necessarily the true social utility maximizer.) ScoreVoting.net/BayRegsFig.html

or i just think having two cases of the same letters is archaic and inefficient, and pointless since spoken words have no case.

i switched to dvorak around 1999. i haven't looked back. it is just so far superior to qwerty, in terms of comfort, that i couldn't imagine it any other way.