YouGov

[CatholicLeft]

Jul 3, 2024 17:43:56 GMT islington said:

Jul 3, 2024 16:01:30 GMT hullenedge said:

The vote shares attached to this are:
Lab 39
Con 22
Ref 15
LD 12
Grn 7

I'm not going to try everyone's patience by explaining yet again why I don't find it credible that these vote shares would generate the distribution of seats in the MRP.

In about 33 hours (give or take) I'll be on here either eating humble pie (which I promise to do if I'm wrong, although I draw the line at a hat), or else saying with unbearable smugness, "Told you so, told you so."

I happen to agree with you. I can see a Labour majority of around 150.

[iain]

Jul 3, 2024 17:43:56 GMT islington said:

Jul 3, 2024 16:01:30 GMT hullenedge said:

The vote shares attached to this are:
Lab 39
Con 22
Ref 15
LD 12
Grn 7

I'm not going to try everyone's patience by explaining yet again why I don't find it credible that these vote shares would generate the distribution of seats in the MRP.

In about 33 hours (give or take) I'll be on here either eating humble pie (which I promise to do if I'm wrong, although I draw the line at a hat), or else saying with unbearable smugness, "Told you so, told you so."

I think you’re engaged in pretty transparent wishcasting when it comes to your predictions of the Lib Dem seat total, and expect us to do far better than in your prediction.

When it comes to this MRP, I can’t disagree with you. Coming even close to 70 seats on these voteshares is exceedingly unlikely.

[islington]

Jul 3, 2024 20:00:52 GMT iain said:

Jul 3, 2024 17:43:56 GMT islington said:

The vote shares attached to this are:
Lab 39
Con 22
Ref 15
LD 12
Grn 7

I'm not going to try everyone's patience by explaining yet again why I don't find it credible that these vote shares would generate the distribution of seats in the MRP.

In about 33 hours (give or take) I'll be on here either eating humble pie (which I promise to do if I'm wrong, although I draw the line at a hat), or else saying with unbearable smugness, "Told you so, told you so."

I think you’re engaged in pretty transparent wishcasting when it comes to your predictions of the Lib Dem seat total, and expect us to do far better than in your prediction.

When it comes to this MRP, I can’t disagree with you. Coming even close to 70 seats on these voteshares is exceedingly unlikely.

Well, you've correctly divined that I'm not exactly, or even approximately, the Lib Dems' no.1 fan (how did I give myself away?). But I'm trying to be objective about their chances.

If you think that the LDs will end up with a vote share of, say, 18 or 20%, then yes, I'd accept that with efficient targeting they might well win 60 or 70 seats, especially with the Tories so weak. But their polling shows no sign of such a striking voting advance: they've nudged up a bit from 9% at the start of the campaign to 11 or 12 now, but assuming that's where they end up tomorrow they will have made little or no advance on their vote share in 2019. I can accept that with tactical voting they might win more seats than last time, but many MRPs suggest they are in line for five or six times as many and I just can't see it.

[Sg1]

Is it possible that the MRP's are expecting unrealistic levels of tactical voting in favour of the left leaning parties?

[Merseymike]

Jul 3, 2024 20:25:14 GMT islington said:

Jul 3, 2024 20:00:52 GMT iain said:

I think you’re engaged in pretty transparent wishcasting when it comes to your predictions of the Lib Dem seat total, and expect us to do far better than in your prediction.

When it comes to this MRP, I can’t disagree with you. Coming even close to 70 seats on these voteshares is exceedingly unlikely.

Well, you've correctly divined that I'm not exactly, or even approximately, the Lib Dems' no.1 fan (how did I give myself away?). But I'm trying to be objective about their chances.

If you think that the LDs will end up with a vote share of, say, 18 or 20%, then yes, I'd accept that with efficient targeting they might well win 60 or 70 seats, especially with the Tories so weak. But their polling shows no sign of such a striking voting advance: they've nudged up a bit from 9% at the start of the campaign to 11 or 12 now, but assuming that's where they end up tomorrow they will have made little or no advance on their vote share in 2019. I can accept that with tactical voting they might win more seats than last time, but many MRPs suggest they are in line for five or six times as many and I just can't see it.

Depends. They can win more Lib Dem seats, but many get virtually none. That's because those MPs are likely to win for Labour.

But some LibDem seats get very few Labour votes. Tactical voters.

In my view, LibDems will mostly get paper candidates. No leaflets.

[iain]

Jul 3, 2024 20:25:14 GMT islington said:

Jul 3, 2024 20:00:52 GMT iain said:

I think you’re engaged in pretty transparent wishcasting when it comes to your predictions of the Lib Dem seat total, and expect us to do far better than in your prediction.

When it comes to this MRP, I can’t disagree with you. Coming even close to 70 seats on these voteshares is exceedingly unlikely.

Well, you've correctly divined that I'm not exactly, or even approximately, the Lib Dems' no.1 fan (how did I give myself away?). But I'm trying to be objective about their chances.

If you think that the LDs will end up with a vote share of, say, 18 or 20%, then yes, I'd accept that with efficient targeting they might well win 60 or 70 seats, especially with the Tories so weak. But their polling shows no sign of such a striking voting advance: they've nudged up a bit from 9% at the start of the campaign to 11 or 12 now, but assuming that's where they end up tomorrow they will have made little or no advance on their vote share in 2019. I can accept that with tactical voting they might win more seats than last time, but many MRPs suggest they are in line for five or six times as many and I just can't see it.

Indeed, 11-12% is little (if any) increase on 2019, but your comparisons with past elections seemingly ignore that the Tories are forecast to receive only 20% or so, a far worse result than even 1997. On that basis, 30 or, on a very good night, 40 seats hardly seems unreasonable despite your protestations.

We are agreed (along with the vast majority of Lib Dems on here!), however, that 60 or 70 seats is unrealistic, and, as you say, we don’t have long to find out if your prediction of 17 or the approximate median of LD posters of ~35 is correct!

[No Offence Alan]

Jul 3, 2024 20:25:14 GMT islington said:

Jul 3, 2024 20:00:52 GMT iain said:

I think you’re engaged in pretty transparent wishcasting when it comes to your predictions of the Lib Dem seat total, and expect us to do far better than in your prediction.

When it comes to this MRP, I can’t disagree with you. Coming even close to 70 seats on these voteshares is exceedingly unlikely.

Well, you've correctly divined that I'm not exactly, or even approximately, the Lib Dems' no.1 fan (how did I give myself away?). But I'm trying to be objective about their chances.

If you think that the LDs will end up with a vote share of, say, 18 or 20%, then yes, I'd accept that with efficient targeting they might well win 60 or 70 seats, especially with the Tories so weak. But their polling shows no sign of such a striking voting advance: they've nudged up a bit from 9% at the start of the campaign to 11 or 12 now, but assuming that's where they end up tomorrow they will have made little or no advance on their vote share in 2019. I can accept that with tactical voting they might win more seats than last time, but many MRPs suggest they are in line for five or six times as many and I just can't see it.

Presumably the models' calculations are that the LDs are down on average, say, 2% in 500 seats but up 10% on average in 100 seats. My LD score in the prediction competition is 41.

[islington]

Jul 3, 2024 23:43:43 GMT No Offence Alan said:

Jul 3, 2024 20:25:14 GMT islington said:

Well, you've correctly divined that I'm not exactly, or even approximately, the Lib Dems' no.1 fan (how did I give myself away?). But I'm trying to be objective about their chances.

If you think that the LDs will end up with a vote share of, say, 18 or 20%, then yes, I'd accept that with efficient targeting they might well win 60 or 70 seats, especially with the Tories so weak. But their polling shows no sign of such a striking voting advance: they've nudged up a bit from 9% at the start of the campaign to 11 or 12 now, but assuming that's where they end up tomorrow they will have made little or no advance on their vote share in 2019. I can accept that with tactical voting they might win more seats than last time, but many MRPs suggest they are in line for five or six times as many and I just can't see it.

Presumably the models' calculations are that the LDs are down on average, say, 2% in 500 seats but up 10% on average in 100 seats. My LD score in the prediction competition is 41.

H'mm. I am sceptical about the degree of targeting you imply but maybe I'm wrong about that. A point in your favour is that my initial estimate was posted at a time when Labour were polling about 43% and this led me to conclude that in seats where the Tory vote collapses the beneficiary would tend to be Labour rather than the Lib Dems. But if Labour end up with 37% (say), then they are less well placed to pick up these seats. They'll still win many of them, though, and a slightly less imposing Labour vote share may help the Tories to hang on to a few. And yes, admittedly hedging my bets here, the Lib Dems probably will win more than my original estimate (17), although I'll be very surprised if it's as many as 41 (unless their vote share turns out to be significantly more than 12%).

After such a wearisome and interminable campaign, it's hard to believe that a mere 24 hours from now we shall have some real results to look at and I'll be on the way to discovering whether I have indeed been talking through my hat all this time.

[greenman]

I have found this site really helpful at You Gov:

yougov.co.uk/topics/politics/trackers/voting-intention

What the daily tracking up to July 3rd shows, not unlike 2015 is that the Conservative vote has been inching up and Reform declining over the last few days.

But in the 2019 EU election one pollster noted that more Green voters were intending to vote than Conservatives.

In this election I think Liberal Democrat and Green voters have reason to vote and given that up to a third of 2019 LD voters have gone over to Labour in some parts of the country I am not surprised to see the LD vote down overall.

But remember 2015 when all those LD voters switched to Conservative, well some of them may be coming back in 2024.

My big question is whether LD and Green voters will switch to Labour in seats where Reform has a chance of winning.

Beyond that some of these LD-Conservative seat battles are very tight races, but I wonder whether Conservatives have the capacity to pull their vote in this election as the wheels may have come off their campaign bus.

That is also going to be the problem for Reform, who according to some do not have good GOTV experience whereas Labour do beyond the three seats You Gov says Reform can win.

[Sg1]

YouGov held up quite well in many respects. It picked up on strong Conservative performances in Ynys Mon, Chingford and Keighley, and potential gains for Labour that weren't on anybody's radar in any serious way. wasn't far off the total number of seats either. I suppose it's very difficult to quantify potential support for independents, which I wouldn't blame them for.

[hullenedge]

YouGov placed 2nd amongst the forecasters.

[Crimson King]

How would we have done?

[michaelarden]

Good to see Electoral Calculus confirmed as a pile of rubbish.

[polupolu]

Jul 7, 2024 16:28:30 GMT michaelarden said:

Good to see Electoral Calculus confirmed as a pile of rubbish.

If you put the UK actual vote share (rounded to 0.1%) for each party into Electoral Calculus, it gives you:

Con 124 vs 121 in reality

Lab 403 vs 412

Lib 68 vs 72

SNP 19 vs 9

Ref 9 vs 5

PC 4 vs 4

Grn 3 vs 4

Adjust for the Scotland-only numbers increases Lab and reduces the SNP and Tory  MPs to get a fraction closer to the actual result.

[iainbhx]

Jul 7, 2024 17:42:00 GMT polupolu said:

Jul 7, 2024 16:28:30 GMT michaelarden said:

Good to see Electoral Calculus confirmed as a pile of rubbish.

If you put the UK actual vote share (rounded to 0.1%) for each party into Electoral Calculus, it gives you:

Con 124 vs 121 in reality

Lab 403 vs 412

Lib 68 vs 72

SNP 19 vs 9

Ref 9 vs 5

PC 4 vs 4

Grn 3 vs 4

Adjust for the Scotland-only numbers increases Lab and reduces the SNP and Tory  MPs to get a fraction closer to the actual result.

But did it do that on the night?

[evergreenadam]

The final YouGov MRP got the Lib Dem seat total exactly right at 72.

However they didn’t end up winning Godalming & Ash as had been predicted by YouGov. So which was the seat they won that wasn’t predicted by YouGov?

[hullenedge]

Jul 9, 2024 19:20:18 GMT evergreenadam said:

The final YouGov MRP got the Lib Dem seat total exactly right at 72.

However they didn’t end up winning Godalming & Ash as had been predicted by YouGov. So which was the seat they won that wasn’t predicted by YouGov?

Mid Dunbartonshire?

[iain]

Jul 9, 2024 19:25:10 GMT hullenedge said:

Jul 9, 2024 19:20:18 GMT evergreenadam said:

The final YouGov MRP got the Lib Dem seat total exactly right at 72.

However they didn’t end up winning Godalming & Ash as had been predicted by YouGov. So which was the seat they won that wasn’t predicted by YouGov?

Mid Dunbartonshire?

Correct, and it also gave us North Dorset but not Inverness.

Astonishingly good hit rate.

[andrewp]

The New statesman are reporting a public preference from July in a hypothetical 2 horse Lab v Con race, of Lab 61% Con 39%. Im not absolutely sure if this is a poll or something else, but the source is YouGov.

But the numbers may make a fair amount of sense. If you assume that most RefUK voters would opt Conservative in that scenario, most Green voters would opt Lab, the Oth- a lot of which is SNP, PC and left wing Independents- goes mostly Lab, and Lib Dem’s maybe 75% Lab 25% Con, then the numbers get to about 61/39

www.newstatesman.com/politics/polling/2024/08/the-end-of-keir-starmers-honeymoon

[The Bishop]

I think that is a regular question that is asked at or around GE day and was referred to previously a few weeks back.

It has been done by at least one pollster since at least the 1983 election when it recorded what remains the best Tory score - with them beating Labour by 59-41. This one equals the previous all time record set by Labour in the 1997 GE (I think it was also 60-40 in 2001)