Votes are rarely actually "wasted", even in FPTP, and people really need to come to a better understanding of that fact because you get the "smartest" people incorrectly thinking their vote is wasted and thus cutting themselves out of the voting pool.
There is a huge difference between 50.1% A, 49.9% B and 85% A, 15% B. One election produces a politician constantly swinging to the center and watching his back for the next election, the other produces one that acts freely on his beliefs. There is a huge difference between 30% A, 40% B, 30% third-party and 10% A, 89% B, and 1% third-party. One clearly tells the dominant A and B parties that the third-party has tapped into something the voters care deeply about but they've failed to serve, and generally at least one, if not both, of the dominant parties will scramble to serve that demographic. The other says that the vast bulk of people are at least mostly satisfied with the choices (which is all you ever can have). It is this constant scrambling to serve dissatisfied third-party voters that creates the two dominant parties in the first place, not some sort of mega-conspiracy.
Even in FPTP, in an iterated voting situation (which is what we have), third party votes in an otherwise uncontested election aren't wasted, they're probably louder on a per-vote basis than the rest of the horde. If you find yourself nearly alone voting Green/Libertarian/Constitution/whatever, that's because most people are not so dissatisfied that they have to vote third party. You're a fringe vote. (So am I.) But it isn't "wasted".
Please note I am only defending my position that votes are not "wasted" in FPTP. I am not defending the proposition that FPTP is optimal. There may be better systems. Most, if not all, voting systems I know of exhibit the property where a vote is never truly "wasted" in an iterated voting situation, so this argument is neither an attack or defense of FPTP since this is not a unique characteristic of any system.
(It is however only one in a pile, by definition of vote. I do fear some people are ultimately complaining that their vote is wasted precisely because we are voting in the first place and not simply taking their opinion as dictatorial writ. There is a certain fundamental dilution-of-voice that you will experience in any voting scheme; arguably, the "big idea" of democracy is precisely that your voice is at least in theory no louder than any other.)
The only people who vote "fringe" in FPTP are people who are so disenfranchised by both major parties that they are interchangeably bad to them, they actually like the one they "know" will win anyway, or alternatively know it's a lost cause, so they take the opportunity to register a protest vote without the risk of changing the actual result to a worse outcome from their perspective.
Everyone else is voting tactically for the lesser of two evils. So not only are votes "wasted" in the sense of being thrown away on parties that have no chance of winning, they're also actively miscounted as support for party A or B, when often they're registering revulsion at the opposite party.
AV isn't perfect but the one thing it does do better than FPTP is allow you to signal your true preference in exactly this manner without that preventing you from expressing which of the major candidates you prefer. It doesn't directly translates your single vote into 1/population worth of representation like PR, because there's only one winner from each region, but it's certainly a reduction in waste and a better reflection of voter desires.
There is a huge difference between 50.1% A, 49.9% B and 85% A, 15% B. One election produces a politician constantly swinging to the center and watching his back for the next election, the other produces one that acts freely on his beliefs. There is a huge difference between 30% A, 40% B, 30% third-party and 10% A, 89% B, and 1% third-party. One clearly tells the dominant A and B parties that the third-party has tapped into something the voters care deeply about but they've failed to serve, and generally at least one, if not both, of the dominant parties will scramble to serve that demographic. The other says that the vast bulk of people are at least mostly satisfied with the choices (which is all you ever can have). It is this constant scrambling to serve dissatisfied third-party voters that creates the two dominant parties in the first place, not some sort of mega-conspiracy.
Even in FPTP, in an iterated voting situation (which is what we have), third party votes in an otherwise uncontested election aren't wasted, they're probably louder on a per-vote basis than the rest of the horde. If you find yourself nearly alone voting Green/Libertarian/Constitution/whatever, that's because most people are not so dissatisfied that they have to vote third party. You're a fringe vote. (So am I.) But it isn't "wasted".
Please note I am only defending my position that votes are not "wasted" in FPTP. I am not defending the proposition that FPTP is optimal. There may be better systems. Most, if not all, voting systems I know of exhibit the property where a vote is never truly "wasted" in an iterated voting situation, so this argument is neither an attack or defense of FPTP since this is not a unique characteristic of any system.
(It is however only one in a pile, by definition of vote. I do fear some people are ultimately complaining that their vote is wasted precisely because we are voting in the first place and not simply taking their opinion as dictatorial writ. There is a certain fundamental dilution-of-voice that you will experience in any voting scheme; arguably, the "big idea" of democracy is precisely that your voice is at least in theory no louder than any other.)