The Vote UK Prediction

[Andrew_S]

As I thought might be the case. the VoteUK prediction figures have hardly changed at all since I last updated the spreadsheet about 3 weeks ago. In fact I think the Lab and Con figures are identical.

Mar 2, 2015 21:59:20 GMT manchesterman said:

So my updated figures for the main parties are now:

Lab 285
Con 275
SNP 34
LD 30

The latest UK vote prediction figures are:


Lab 286
Con 273
SNP 31
LD 33

Assuming they are both more or less accurate (and nothing major changes in the next 2 months), we would be heading for a situation whereby a Con-LD or a Lab-LD or a Lab-SNP arrangement would not be enough for a majority.

[woollyliberal]

Mar 2, 2015 19:06:53 GMT manchesterman said:

Here's a bit of work I did over the weekend. Basically used Poll of Polls figures, then added in an incumbency bias. Then, as it was far from detailed analysis, I left out the N. Ireland seats, gave UKIP their 2 current seats + Thanet South (just to make it a little more realistic). Other than that, I havent factored in any regional differences, local issues, candidates issues etc and therefore is not a prediction as such, but (in Peter Snow's words) is "just a bit of fun".

docs.google.com/spreadsheets/d/19rpGTdG__XdasJxiZyMv3FAEeHctj9O94t0lZxFqWw4/edit#gid=1793099435

Interesting to note how similar the results are to the Vote UK prediction results mentioned elsewhere in this thread. Very similar indeed - although a few seats differ here and there as you HAVE taken into account constituency factors/issues etc.

The thrust of both projections is that neither large party may be able to form a government with either JUST the Lib Dems or the SNP.

Interesting times ahead!

I like that you're predicting nevative percentages in some seats. Now that's a real swing away.

[manchesterman]

Haha. I did try to edit those seats to zero but I may have missed a few due to lack of diligence

[manchesterman]

Maps relating to my extrapolation posted above..

[manchesterman]

[bossmark]

I predict that Labour will gain my seat of Elmet and Rothwell.

[iain]

Obviously you said that it wasn't precise on every seat, I think this is definitely true to the LD seats. Though Labour gaining Southport is an odd one

I think you're off in Scotland particularly though.

[Pete Whitehead]

Mar 8, 2015 12:46:01 GMT manchesterman said:

How do you get Labour gaining Gravesham?

[No Offence Alan]

Eastleigh Blue?

[Deleted]

We're quite right to point out any errors on manchesterman 's maps upthread, but we're also at risk of sounding ungrateful for something that would have involved a lot of time and effort, so I would like to thank him.  I'm sure it won't take too long to rectify.

[bolbridge]

Mar 2, 2015 21:59:20 GMT manchesterman said:

So my updated figures for the main parties are now:

Lab 285
Con 275
SNP 34
LD 30

The latest UK vote prediction figures are:


Lab 286
Con 273
SNP 31
LD 33

Assuming they are both more or less accurate (and nothing major changes in the next 2 months), we would be heading for a situation whereby a Con-LD or a Lab-LD or a Lab-SNP arrangement would not be enough for a majority.

Your figures: Lab-SDLP-SNP on 322 where 323 is effectively a majority (given the speaker and SInn Fein). 3 PC, 1 Green gives Labour enough wiggle room that they could do a minority and play the Lib Dems and SNP off against each other - I imagine we would hold out for opposition and watch the SNP screw Labour initially.

The second scenario would mean 320 for be same combination - but Plaid would probably only deal in tandem with the SNP so this would actually help Plaid. I would imagine another minority Labour in this case.

In either case Labour could probably work with the DUP at times - but I don't know how this would affect their relationship with the SDLP.

[Andrew_S]

I'm halfway through updating the spreadsheet. (It's taken a couple of hours so far, mainly because of the old laptop I'm using at the moment).

[manchesterman]

Sorry for the handful of errors you diligent folk have discovered.

I will try to correct a few in the next few days. Southport and Gravesham were clear oversights! I did actually have the Tories winning Eastleigh though!

[Andrew_S]

I've updated my personal spreadsheet and despite lots of new votes over the last few weeks hardly any seats "changed hands". Thurrock is currently UKIP instead of Labour and Norwich North is Labour instead of Conservative. A couple of seats in Scotland changed hands as well.

I haven't been able to update the online version of the spreadsheet because my laptop crashed just before I was about to do so and its still in a state of suspended animation. (It's very difficult to do spreadsheet work on an iPad which is what I'm using at the moment).

[Robert Waller]

For comparison, have we had this 'panel of experts' prediction anywhere on Vote UK?
www.psa.ac.uk/insight-plus/blog/what-about-election-experts

Apart from not adding up to 650 and apparently not understanding Northern Ireland, I myself think they're way over on Labour. If the SNP really do break through, I think Labour may be lucky to end up with more seats than they have at present.

[greenchristian]

Nate Silver has announced that he's going to be doing a prediction, though he says it's in conjunction with electionforecast.co.uk, and he's not released any figures yet.

[East Anglian Lefty]

Mar 9, 2015 22:43:25 GMT greenchristian said:

Nate Silver has announced that he's going to be doing a prediction, though he says it's in conjunction with electionforecast.co.uk, and he's not released any figures yet.

I predict it'll do comfortably worse than UNS, and probably worse than most other projections to boot.

[Andrew_S]

I've updated the spreadsheet:
docs.google.com/spreadsheets/d/10RLXcjEC53skefzikBQ9GC9CzMMqDdGisXxPzQ02ayM/edit?pli=1#gid=0

Latest projection:

Lab 286
Con 273
LD 34
SNP 29
UKIP 5
PC 3
Green 1
Spkr 1
NI 18

The previous figures from 7th February were exactly the same for the two main parties. LDs were on 33, SNP were on 31, UKIP were on 4.

[Andrew_S]

Seats where the gap between the first and second-placed parties is 3 votes or less:

Norwich North, Stevenage, Thurrock.
Northampton North.
Stockton South.
City of Chester, Wirral West.
Crawley, Portsmouth South.
Kingswood, St Austell & Newquay.
Dudley South, Halesowen & Rowley Regis.
Bradford West.
Airdrie & Shotts, Berwickshire Roxburgh & Selkirk, Caithness Sutherland & Easter Ross, Glenrothes, Inverclyde, Kilmarnock & Loudoun, Lanark & Hamilton East, Linlithgow & East Falkirk, North East Fife, Paisley & Renfrewshire North, Stirling, West Dunbartonshire.
Belfast East.

No seats in London or Wales are "close" using this definition.

[Deleted]

Mar 9, 2015 21:08:24 GMT Robert Waller said:

For comparison, have we had this 'panel of experts' prediction anywhere on Vote UK?
www.psa.ac.uk/insight-plus/blog/what-about-election-experts

Apart from not adding up to 650 and apparently not understanding Northern Ireland, I myself think they're way over on Labour. If the SNP really do break through, I think Labour may be lucky to end up with more seats than they have at present.

The correct 'academic' answers to the question of party vote shares and seat counts are respectively "not enough data to say" and "not enough data to say". What they are doing here is playing parlour games. Nothing wrong with that, but let's not pretend it is some kind of academic exercise.