- Following the Rochester and Strood by election, Conservative chances of remaining in coalition diminished, while their slim chance of a majority stayed the same. Labour have had a modest increase in their chances, but nothing that's outside of recent movement. As mentioned before, UKIP are still predicted to lose their by-election gains in the general election.
As a little demonstration, I've also run a simulation of the completely unlikely case that we have a sudden snap election this week. This is purely to demonstrate the difference between what the model thinks future vote share will be and what the current polling is. It's quite a minor difference in vote share, but in election outcome it's much larger, as it tips Labour over into being odds on favourite to form a government, and over a third of simulations result in a labour Majority.
The difference in the prediction is mainly dictated by reversion to mean, as consumer confidence remains relatively high, while the momentum of the two largest parties remains either flat or slightly negative. Economic shocks would be very bad for the Conservatives chances in the model, but they wouldn't really benefit Labour due to a lack of positive momentum. UKIP has the most to gain if revert to mean eases off, but their support seems to have fallen back a little, and they may well have peaked far too early.
Source UK Election trend