Martin Kettle's piece on political reform starts off sensibly enough, noting that the window for action is really pretty small as, after the election, the new government will find bigger fish to fry. But soon enough, we get on to electoral reform and discover once again that there is no political problem that a new voting system won't solve.
In this case, the "main advantage" of the AV system is that it will ensure MPs have a true majority of votes. Kettle doesn't actually say why this is an advantage; I'm going to guess that the idea is that it would encourage MPs to worry more about keeping all their constituents happy. This is probably true, in that MPs who rely on second or third preferences to get elected will want to maintain a minimum level of acceptability. That's why it's a bad idea.
MPs, frankly, should be prepared to make unpopular decisions. AV puts a campaigning MP in the position of persuading her 1st preference supporters that she's true to the ideals of her party, and her opponents' 1st preference supporters that she's a perfectly acceptable alternative to their party. It's a balancing act, with the concomitant risk of falling between two stools. Proponents will argue that this leads to government by consensus, which I don't doubt. But it's not a consensus based on winning the argument, just on finding an innocuous middle-ground: the consensus of the lowest common denominator.