Let's look forward to a potential election in 2015.
In your constituency the result is counted and it ends as thus..
Lib Dems 9%
Happy with this result? Over half have voted for the Tory candidate, and he's elected. Good stuff? Looking at it as a FPTP result, it's clear it's a fair win for the Tory MP.
If this is the case, then how does the presence of other rounds before it, should it also be the AV result, make any difference? What if the round before the final round above was...
Lib Dems 8%
And how about the round before this it looked like this?
Lib Dems 8%
Are we honestly so fickle that we would accept the result at the top of the page if it were FPTP, where voters have decided to vote Tory in the absence of their own candidate, but won't accept the same result achieved through a different system?
EDIT: I'm just going to elaborate on my thinking, after a conversation with @SohoPolitico on twitter.
The reason I bring this example up is it's important to understand that preferences that people place explicitly under AV are the same preferences that they decide to use implicitly during FPTP.
Take the top result on this page between three candidates. If it were FPTP it would be accepted as a Tory win by everyone, perhaps as a clear mandate. The reality would be that the vote is made up by people that have used their own second and third preference choices as though it were a first preference because they don't want Labour to win.
The result that would take place under FPTP would not be transparent, we wouldn't know if the Tories had a real majority, or if the vote was more about not wanting Labour while *really* wanting someone else. But the one universal truth between both that result as FPTP, and AV, is the same voter in both scenario's feels strongly about voting for someone other than Labour and has every right to express that preference.
As soon as you understand that how people feel doesn't change between the two results, regardless of how the result is reached...you should also understand that comparing people's preference level to each other is irrelevant. We don't measure inter-personal preferencing under FPTP, it's implicit, but just because it's out in the open doesn't mean it's relevant to "who is more popular" either. Preferences only exist to allow the system to work.
If the system was Run-Off, where there is a re-vote between rounds, we wouldn't use preferences...we wouldn't need to. The system is functionally identical except for this fact...preferences therefore only exist to allow vote counters to carry out each round without reapplying for people's opinion.
So with preferences being irrelevant to the validity of the "popularity" of the result, it just comes down to that result. If you accept the top result on this page is a victory for the Tories then you also accept that by adding more people in to the mix the Tory hasn't got less popular in a head to head with Labour.
The only thing that has changed as you go down the "rounds" is that people are splitting their vote...that the Left leaning people in the constituency are all huddled around Labour while an increase in right wing candidates spoils the vote for the Tories.
This is why AV is a fairer system, we can see where the vote is being split, we can rectify it by ensuring that we understand who the most popular person is in a final head-to-head (or head-to-heads), with an accurate level of information as to the specific type of policies that the winner should be taking in to consideration if they are to be truly representative of their constituency.
AV... more representative of overall views
AV... more transparent for telling the core priorities of electorate
AV... Disallows the logistical act of vote splitting to enable an unpopular candidate to win.
This is why it is a no brainer to vote to change to AV this May.