By Michael Thrasher, Sky's Election Analyst
With 100 days to go until the General Election, Sky News is forecasting another hung parliament - with Labour rather than the Conservatives as the largest party.
Typical: you wait years for a hung parliament and then two come along together.
This forecast will change because electoral opinion is changing. Gone are the days when parties could rely upon a solid core of support and concentrate instead on wooing the "floating" voters.
Over the past two years, former Tories have separated from the party, tempted by the promises of UKIP to curb immigration and abandon the European Union. It is 100 days and counting before separation turns into divorce.
Labour, which recovered rapidly from its 2010 defeat because of Liberal Democrat voters defecting, has squandered that inheritance. Deemed too left-wing or too right-wing according to which branch of the party has a public platform, it is being outmanoeuvred by the radical Greens in England and by the SNP in Scotland.
The electoral arithmetic is clear. Understanding what might happen on 7 May requires us to reset the parliamentary numbers to what they were after the last election.
The Conservatives, with 306 seats, fell 20 seats short of an overall majority, despite staging one of its best-ever election performances.
Labour finished with 258 seats, and required 68 gains for a victory.
:: Sky looks at the 150 seats that could play a deciding role in May's General Election. Click here for the link to the In The Margins console.
Election watchers normally talk about swing - the movement of voters from one party to another across consecutive elections. The line for next May is pretty well-rehearsed.
The Conservatives require a swing of 2% from Labour (or an 11-point lead in the popular vote) for an overall majority.
Meanwhile, Labour overtakes the Conservatives with a 2% swing in its favour, and then crosses the finishing line with a 5% swing (a three-point lead).
The national polls currently show that Conservative support has fallen by six points, while Labour has enjoyed a rise of three points. Assuming a national uniform swing, Labour would be just short of a majority.
But the 2015 election is so unlike previous elections that such forecasting conventions should be abandoned.
National polls are now too blunt to capture the undercurrents of electoral opinion. They fail to reflect the turbulence in Scotland following the referendum vote, differences across the English regions, and politics in individual constituencies.
Last September, 45% of voters in Scotland supported independence. The latest Scottish-only polls shows most of these will stick with the SNP for the Westminster election.
Meanwhile, the majority that voted to stay with the union have scattered among other parties.
All of this is good news for the SNP, which could win 53 seats, and disaster for Labour, projected to lose 36 seats.
There are differences in the English regions, too. The UKIP threat to the Conservatives is concentrated down the east coast, and into Essex and Kent particularly.
While Conservative hopes are pinned on banging the drum of economic recovery, its sound is muffled in the Midlands and the north, where 38 of its most marginal seats are situated.
Opposition parties would hope to tap into such scepticism, but these voters are unimpressed with Labour's leader and his party's message.
General elections are ultimately decided not by the national vote, but by results in 650 separate constituencies. The contrast here between UKIP and the Liberal Democrats reinforces this point.
Farage's party is generally more popular than Clegg's, but vote distribution is everything. UKIP may win Clacton and Rochester again, but the Liberal Democrats could have 10 times that number of seats despite polling fewer votes nationally.
A hung parliament is therefore the most likely outcome of this chaos, but the 100 days after the election look equally unpredictable.
