Monday, 13 April 2015

UK GENERAL ELECTION PREDICTION BY UK ELECTION TREND

This is a moderately significant update, and just to make sure you take it seriously I remind you it's far too late for this to be an April fools. As the model ticks forward each day closer to the election, the weak reversion to mean is applied, and old polling is aged out of the recent average. Prior to today, and as I noted in previous posts, the model has been pessimistic about the SNP's ability to get a landslide in Scotland. This was in parts due to seeing a small pre-debate slide in the SNP's polling, applying a weak reversion to mean, and the results seen when doing a monte-carlo prediction. Because a bulk of the SNP's target seats were in variance range of being defended, when the simulations were taken in sum a number of seats were expected to be defended. No seats in particular were really identified as being particularly defendable. Instead in any particular simulation run of the thousands performed, some of those target seats would be defended, and in enough simulation runs in total to move the prediction.

However, this was Friday's run of the model. The simulations have now reached a tipping point where the bulk of SNP target seats moved out of a variance range where they could be defended.

The result is that in the majority of simulations, the SNP win a landslide in Scotland. The model now has a staggering 70% probability that the SNP will hold the balance of power in a hung parliament, and based on statements made by the party leaders that can be interpreted as a Labour minority government. There is only a total 4% of simulations showing any party having a majority, with Labour reduced to the same level as the Conservatives. However, they still retain a moderate edge over the Conservatives on being able to form coalition with the LibDems, at 13% to 5%. With the possibility of UKIP-Conservative coalition still being all but ruled out as they fail to establish general election wins in any simulation run at all.

I say this is only "moderately" important a change, because the model had already predicted that the SNP would hold the balance of power. This "merely" moves it into that much higher a probability with a likely landslide north of the border.

No comments:

Post a comment

Comment is open to all feel free to link to this blog.