Election Predictions

[drplokta]

Anyone care to have a guess at the percentage breakdown of the national election result (excluding Northern Ireland constituencies)? No more new entries after 10pm on Thursday evening, which will be exactly a week before the polling stations close and we start to get exit polls. The winner will be the person who has the lowest total difference between their guess and the actual result. The smaller parties (SNP, Plaid Cymru, UKIP, BNP, Respect, etc.) are all bundled together under "Other". I recommend that your choices add up to 100%, but it's not compulsory.

My guess is Conservatives 36%, Labour 29%, Lib Dems 25%, Other 10%. That would be a hung Parliament unless something very odd happens at the constituency level, with the Conservatives having a few more seats than Labour.

Comments

[evil-mogwai.livejournal.com]

This will probably be comically bad:

Lib Dems 29% Conservatives 28% Labour 27% Other 16%


It seems that most people who have not been strictly aligned with either Labour or Conservative seem to me more interested in either voting Lib Dem "for a change" or want to vote UKIP.

[ffutures.livejournal.com]

Lib dems 32%, conservative and labour around 30%, the rest 10%, but still a conservative victory due to the crappy system.

[andrewducker]

That would be very like to be a Labour victory - still hung, but only very slightly.

[coalescent.livejournal.com]

Conservatives 33%, Labour 29%, Lib Dems 28%, Other 10%. That is, I think the final vote will be quite closely bunched between the three parties, and Lib Dem support will hold up better than you're guessing, but not enough to overtake Labour.

[ang-grrr.livejournal.com]

C 33, L 30, LD 28, Other 9.

I think Lab/LD waverers like me will panic come the election and the percentage vote for the LD will dip slightly. I'm expecting a number of UKIPers and the like to shift to Con to try and get Labour out.

[ang-grrr.livejournal.com]

And now I'm wondering if I should revise this given Gordon Brown has called a labour voter a bigot...

[andrewducker]

I take it you mean overall share of the vote, not of the seats?

In that case:
Con: 35%, Labour: 30%, LD: 26%, Other: 9%

Not dramatically different to you. I expect Lab/Con to get a small boost over current polling, but not much.

[megamole.livejournal.com]

Indeed. Would be very interesting to see comparison of overall voting % to seats won. I'm convinced we need reform of the system, but am not a fan of pure PR.

[andrewducker]

You can have a play with the Seat Calculator here, although it's obviously only an estimate.

[evil-mogwai.livejournal.com]

The seat calculator suggests that the Lib Dems would need 40% to get 243 seats when Labour would only need 24.4% and the Conservatives 30.3%!

IS our system really that broken?

[drplokta]

Yes, it really is. It favours parties who have regional concentrations of strength -- the LidDems do badly, because they perform at similar levels everywhere.

[andrewducker]

Sadly...

[childeric.livejournal.com]

Con 37
Lab 26
LD 28
Other 9

What about seats, or is that too hard to call?

[gaspodia.livejournal.com]

Labour 33%, Conservative 34% Lib Dem 24% and Other 9%

[pigeonhed.livejournal.com]

Labour 33%
Conservative 31%
Lib Dem 27%
Other 9%

I actually expect Labour to break 300 seats but not have an overall majority, something around 315 would be good. It means that other parties have influence but not unduly so. Where more than about 20 others are needed to pass a vote is where things get messy I think.

[bob]

30% con 30% labour 30% libdem 10% other

[bohemiancoast.livejournal.com]

32% Tory, 31% Labour, 29% LibDem, 8% Other

[cobrabay.livejournal.com]

34% Tory, 29% Labour, 27% LD, 10% Other

[purplecthulhu.livejournal.com]

C 33; LD:29; L: 28; O: 10

I think Gordon's gaffe is going to get a lot of airplay. The news media have been looking for the 'gaffe of the election' and, in the absence of anything else, I think they just found it. This will cost Labour dear.

[hal-obrien.livejournal.com]

From Nate Silver, who was far and away the most accurate predictor of the 2008 US Electoral College:

"How about a scenario in which the Liberal Democrats do really, really well? Suppose that they get 20 percent of Labour's voters and 20 percent of Conservatives, as well 10 percent from Great Britain based 'others'. We'll also give an additional 10 percent of Labour's 2005 vote to Conservatives. This would imply that LibDems wind up with 36.4 percent of the vote nationwide..."

That sounds plausible to me in the current circumstances. Per Silver, it yields LibDem 36%, Tory 29%, Labour 25%, and Other 10%.

The "Voter In the Street" interview I've heard that was reallly interesting went like this: "I'm tired of the Labour govt, but I've never voted Tory in my life (and won't). So I'm going to vote LibDem..." I just don't get the impression the Tories have convinced anyone aside from their base that they're human beings, and I think that's going to be a widespread attitude. I also think there are a non-zero number of Tories who are "Not-Labour," rather than Tory as such. The more plausible the LibDems become, the more they'll get a virtuous cycle in swing.

At least, that's the view from this ignorant Amurrican.

[the-gardener.livejournal.com]

Silver's poll analysis was covered in today's Guardian. It all sounds extremely plausible and sensible to me; I have argued for some time that polls which translate national percentages into uniform swings are nonsense because they completely ignore the individual factors which influence the results in each constituency -- whether the incumbent has a personal following, voters' experience of fluctuations in hospital waiting times, factory closures or expansions which affect the local jobless figures, even what the council has been doing about waste collection and parking.

The bottom line, I think, is that -- once the numbers have been crunched and the voting results (whatever they actually are) have been analysed to within an inch of their lives -- UK pollsters will be beating as path to Silver's door after 6 May. And rightly so.

[pierre-fermat.livejournal.com]

Con 35 Lab 28 Lib 26 Oth 11

Giving the Tories more seats but not a majority.

[hal-obrien.livejournal.com]

"Giving the Tories more seats but not a majority"

Nate Silver disagrees, with that kind of split.

"What if the Conservatives do really well instead? (There is some notion that pre-election polls tend to underestimate Conservatives' standing, although I have not investigated this myself). If, for example, conservatives were to get about 37 percent of the vote with Labour and LibDems both between 26-27, our model shows them in control of Parliament with 333 seats, whereas uniform swing gives them only 304."

Again, I give Silver a lot of credit because he has a track record of being accurate.

[pierre-fermat.livejournal.com]

He doesn't seem to have considered that more people are prepared to vote tactically against the Conservatives than against Labour. If the uniform swing machines are wrong it is probably in underestimating the seats that will be won by the Lib Dems and Others. Both Labour and the Tories will lose seats that they wouldn't have on a uniform swing. Labour have more seats to lose but the Conservatives are seen as the real enemy by many Lab and Lib voters who will vote for whichever party is most likely to keep the Tory out.

In 30 hours time I may have been proved very wrong...

[hal-obrien.livejournal.com]

That appears to only result in the strong LibDem showing I allude to upthread.

This is a moving target. Here's his most recent projection, unlike my original pointer, which discussed a number of possibilities. (I keep including links to my sources in the naive hope readers will use them, despite all evidence otherwise. This leads to behavior like saying, "He doesn't seem to have considered..." when he did, it's just I didn't quote that part, because again, I had the hope one would click through if one was interested, and not rely solely on my {obviously} abridged representation of his position.)

[stevegreen.livejournal.com]

Are your percentages for votes cast or seats secured?

[the-gardener.livejournal.com]

Has to be for votes cast, surely. The constituencies are so gerrymandered* that trying to derive seats actually won from votes cast is next to impossible.

* Which might be a good reason for supporting Cameron's claimed desire to reduce the number of MPs and even out the size of the constituencies, did one not suspect that he wishes to gerrymander them in his own way -- uniting chunks of cities which vote Labour with larger chunks of rural hinterland which vote Conservative*, etc. etc..

* A rural hinterland which even then thinks Margaret Thatcher was a bit of a communist. Bloody farmers.

[drplokta]

Votes.

[del-c.livejournal.com]

I think you should score according to root-sum-square of differences rather than the sum of differences. That will better identify who gets closest to the point in four-dimensional space represented by the actual result.

[the-gardener.livejournal.com]

I'm not prepared to give percentages at this point, but I expect the Lib Dems to sink back as Der Tag approaches, and Labour to recover slightly....but not enough to prevent the Conservatives emerging as the largest party. The Tories may not have an overall majority, but if they can stitch up some agreement with the Northern Ireland parties (who historically are closer to the Tories than anyone else) I expect them to be forming our next government in the week commencing 10 May.