A change in how I'm reporting the simulated election results in the graph, by separating out the SNP's seats. In addition, the assumption will now be that the SNP will support a Labour minority government rather than go into coalition.
a week where the campaign heated up about Tax avoidance, Labour has
gained a clear lead in the polling average. However, the model still
assumes the result will tighten up again. This is not due to the revert
to mean assumption, which is making increasingly smaller adjustments to
the vote shares as election day grows closer. Instead it's due to the
model looking at recent trends and assuming that Labour still has a weak
momentum. Should Labour be able to sustain gains from the campaign,
that will change as momentum swings in Labour's direction.
For comparison purposes, I've run a "Nowcast" showing what the
approximated result of an election conducted yesterday would have been.
The chances of an outright Labour majority are greatly increased, and a
significant increase to the chance of coalition, but only a modest
increase to the over all chance of a Labour government of any kind.