It's just as well Iain Dale qualified his prediction of the general election in Scotland by noting that "my knowledge of Scottish seats is not the same as my knowledge of English seats", because two aspects of it look distinctly difficult to justify - the suggestion that the SNP will lose two seats to the Conservatives (presumably Angus and Perth & North Perthshire) and, perhaps even more so, that the Liberal Democrats will reach a new high watermark of fifteen seats. There's nothing complicated in explaining why these predictions are, in Sir Humphrey terms, somewhat "courageous". Firstly, to lose seats to the Conservative party, the SNP would have be to suffering a net swing to the Tories. I have the slight suspicion that Iain is making the schoolboy error of thinking that because the majority in those seats is small, and because he expects the Tory vote to rise, it therefore follows that those seats will fall. But the snag is that all the expectations are that the SNP vote will rise as well, by at least as much - and, let's be frank, probably a good deal more - than the Tories. That would suggest a net swing from the Tories to the SNP, which could even mean increased majorities for the Nationalists in those two marginal seats. The one remaining hope for the Tories in such circumstances is that the nature of the two-horse fight in those constituencies could help them buck the nationwide trend - but is there any reason to suppose it would? My suspicion is that the opposite may happen - that any suggestion of potential Tory gains will cause Labour and Liberal Democrat voters to switch tactically to the SNP.
But for all that I've just said, at least Dale could make a case, however dubious, for that part of the prediction. However, on the Liberal Democrats winning fifteen seats point, words fail me. The Liberal Democrats won 11 seats last time round on the basis of 23% of the vote - they will be extraordinarily fortunate to reach even a 16% or 17% share at the forthcoming election. OK, first-past-the-post sometimes produces quirks, but it would require tactical voting on an industrial scale to get the Lib Dems anywhere close to fifteen seats in these circumstances.
Iain also asks where Alex Salmond thinks his target of 20 seats is going to come from. Again, there's nothing complicated about this - the SNP need to be leading or close to leading in the popular vote to have a chance of attaining that goal. It's of course perfectly legitimate for critics to question whether they can do so, but what is rather silly is to imply as Iain seems to be that Salmond is suggesting that the target can be achieved on a more modest swing, which he plainly isn't. Moreover, there's no mystery as to which 20 seats these would be - Iain could just have pumped the relevant figures into a UNS swing predictor to have gained a rough idea.