UPDATED 0755

Clegg wins by three pollsters to two

Clearly there will be great relief in the blue camp about the outcome last night although, as can be seen, the message from the pollsters in their instant surveys was mixed. In terms of straight wins it was Clegg 3 Cameron 2 with Brown on zero although his performance was seen as being much better than last week.

These are, of course, quickie reactions and will frame this morning’s coverage. The big question is what this will do to the voting intention polls and we’ll need to wait until Saturday before getting a full range.

The Brown camp will be pleased with their man’s performance though his denial on what is in party leaflets on pensioner bus passes might be something that he’ll regret - for it does not take much for opponents to show that his statement was untrue.

The SNP was quick to circulate copies of the Labour leaflets from his own constituency and, no doubt, there’ll be spats for a few hours.

Mike Smithson

Is now the time to do some serious spread betting?

There’s a sub-set of data in the PB/Angus Reid general election polling series that looks at what is happening in the 62 seats which the Lib Dems won at the 2005 general election. Inevitably given that this represents only about one in ten of all GB seats the sub-sample is small.

But the numbers coming from surveys since the debate show that Clegg’s party is doing enormously well in the seats that it holds. Last night’s poll had the split LD 44%: CON 23%: LAB 19%.

This supports what was being shown in pre-debate polling that the Lib Dems are doing well in the seats that they are defending. The chances are, surely, that it will keep almost all of them.

So if you are entering into the high-risk arena of spread-betting on commons seats then you can be relatively confident that Clegg’s party is not going to drop below about 60. If you bought at the current Sporting Index level of 86 seats then, arguably, your maximum losses would be about 26 seats.

Thus a bet of £10 a seat would expose you to a down-side risk of £260. Set that against the potential upside. Thus taking last night’s YouGov poll fed into the Well UKPR seat calculator and you get a 143 seat projection - or a potential gain of 67 or £670 for a £10 a seat bet.

The question is whether the Lib Dems are going to stay up at these dizzy heights or will the bubble burst? Well it hasn’t so far and the surge continues.

All the polls where the fieldwork was completed yesterday had very good numbers for Clegg’s team. The ComRes roly poly poll was slightly earlier and based on the aggregation of samples of 500 from each of Sunday and Monday.

One of the great things about spread betting on commons seats is that if you see your choice fading then you close the bet down at whatever the current sell price is.

Mike Smithson

But the Labour still have a double digit deficit

There’s a new PB/Angus Reid poll just out and the figures are above. Polling started yesterday and went on until this afternoon - and was almost all carried out after the election was declared by Mr. Brown yesterday.

The comparisons are with the weekend’s AR poll for the Sunday Express.

The numbers are good for the Lib Dems who for the next four weeks will benefit from the broadcasting rules that give them coverage at almost the same level as the two main parties.

Support for “others” remains at a high level though that could reduce sharply as we get close to May 6th. The AR voting intention question will be tailored so that respondents are only asked about parties standing in their seat.

The data should be out later this evening and a link posted here.

Mike Smithson

And others drop sharply as well

With the election only days away from being called there’s a boost for the Lib Dems in the latest PB/Angus Reid poll.

The entire field-work took place yesterday and today and so was completed after Monday’s Channel 4 “Ask the Chancellors” debate.

A big feature in this latest poll is a sharp decline in the proportion going to others. This follows a period when AR was recording the biggest shares - that is now down to one of the lowest.

Labour also move up to their highest recorded level since AR started their UK polls last October.

In the run-up to the election AR have made a change to their geographical sampling approach. This is how they describe it:-

“……we believe that one of the most important ingredients of success in online polls is rigorous geographic sampling. This means not simply ensuring we have the appropriate number of respondents in a broad region, like London, but that we the right number of people in each part of London from Tower Hamlets to Croydon. In order to achieve that we have divided the 632 constituencies in Great Britain into 142 “Super Constituencies.”

Each of these “Super Constituencies” is made up of about four or five constituencies grouped together based on a combination of geography, demographics, income and political leaning. This approach means that no micro-political culture is missed, leading to more accurate results.”

This is how AR use operate their campaign polls in Canada where the firm has had great success.

I’ll update this post with the link to the detailed data when that becomes available later tonight.

Mike Smithson

Update (2045): Andy from Angus Reid has confirmed that the figures originally released hadn’t been fully adjusted.  The correct figures are now in the table.  David Herdson