Over the past couple of weeks, I’ve spent quite a bit of time examining the electoral landscapes facing the Scottish and Welsh Tories in this general election.

Each received an entry in our comprehensive battlegrounds series, whilst this column has looked at what YouGov’s first MRP poll augured for the Party and, drawing on that, the question of whether the Welsh Conservatives could get their campaign back on track.

Yet since then there’s been a final round of fresh evidence (or omens, depending on your faith in polls), and it once again poses interesting questions.

(As a lot of polling is conducted on a Great Britain basis which excludes Ulster, I won’t cover it here. Our battlegrounds profile for Northern Ireland, which includes a link to the latest LucidTalk poll of the Province, was published yesterday.)

Is something afoot in Scotland?

In my previous column on YouGov’s first MRP poll, one point I kept coming back to was how different was the story it told both from general expectations before the campaign began and from on-the-ground feedback from activists.

The same is true for the second one. Despite a general sense that the wind is in the sails of the Scottish Conservatives, YouGov has them going backwards.

Where in the first poll they were on track to hold 11 of their current haul of 13 seats, it is now only eight. On top of losing Stirling and East Renfrewshire – and the latter is widely viewed as an implausible win for the SNP – they are now forecast to lose Angus, Gordon, and Ochil & South Perthshire (confusingly the Times graphic has the SNP winning Moray, but that appears to be a glitch). These losses are partially offset by a predicted gain from the SNP in Lanark & Hamilton East.

Meanwhile Scottish Labour, which just a couple of weeks ago were staring down the barrel of a near-wipeout, are forecast to retain five of their current haul of seven seats, including fending off a rumoured Tory upset in East Lothian.

This conflict between YouGov’s forecasts and the wider narrative of the campaign could mean that the pollster has identified trends which have flown almost entirely beneath everyone else’s radar. Or it could mean that they have mis-calibrated their model and got it wrong – we’ll find out tomorrow.

If they are right, it will pose hard questions both to campaigners and commentators – from many parties – about our ability to read a live election campaign, not to mention trying to work out where the Labour recovery came from and why the Conservatives slipped back in the closing weeks of the campaign.

Conflicting auguries for the Welsh Tories

In Wales, the second YouGov report is also disappointing reading, albeit less so. The Conservatives are still forecast to win back their by-election loss of Brecon & Radnor, reclaim Vale of Clwyd, and make an historic advance by winning Wrexham.

But where two weeks ago they were apparently on track to win Anglesey (‘Ynys Môn’) for the first time since the mid-Eighties, the MRP odds now have them in third. None of the other gains they were anticipating to make from Labour at the start of the race are tipped to fall – although they are neck-and-neck in Delyn and just a few points short in both Bridgend and Alyn & Deeside.

This is a strong level of continuity with the previous poll, and if true poses the questions I raised in last week’s column. But it is not the final word on the subject.

Step forward Professor Roger Awan-Scully, of Cardiff University, and his Welsh Political Barometer poll. This is the main, regularly-conducted Welsh opinion poll and it ought to be familiar to readers of this column.

According to him, Labour are on track for an “historic shock” in the Principality. Whilst the Opposition’s vote share has recovered to 40 per cent, he found the Tories on their heels at 37 per cent. This would apparently be the strongest Conservative vote share in Wales since 1900 – before the universal franchise.

Were this to be true, tomorrow night would play out very differently. The Party would not only pick up Brecon & Radnor, Vale of Clwyd, and Wrexham but also Alyn and Deeside, Bridgend, Cardiff North, Clwyd South, Delyn, and Gower. Meanwhile Mark Williams would take back Ceredigion from Plaid Cymru, which YouGov thinks is deeply improbable.

Such results would cut Labour’s representation to 20, just half of Welsh seats, whilst the Conservative would wrack up a better result than their previous high-tide mark at the 1983 election. By contrast the last Welsh-specific poll, on November 25, forecast Labour to lose just four seats.

Once again, we have a fascinating conflict between YouGov’s model and another source, in this case the well-established Welsh Political Barometer. Which is right? It isn’t just a fascinating academic question (although it is that): the MRP model has played a very prominent role in shaping public understanding of the campaign. If parties have based strategic choices on it, and it turns out to be wrong, that could have profound consequences for the result of the election.