The Bluffer's Election Night Handbook

By Adam Ludlow, Senior Consultant

If you’re making the effort to stay up tonight, we’ve compiled a guide of the trends to look out for, what particular results mean and, if you’re really looking to impress, how to do your own back-of-the-envelope seat projections based on early results.


One of the easiest ways to get an early, results-based indication of the eventual “winner” is to look at bellwether seats that have historically gone with the eventual victor. Unfortunately, options are fairly limited this time around. The traditional British bellwether, Basildon, was a victim of boundary review before the last election, and now includes the heavily Conservative Billericay, meaning that it is no longer really in contention. The same fate also befell longstanding bellwether Bristol North West.

Loughborough, another traditional bellwether, is slightly affected this time by having Education Secretary Nicky Morgan as the incumbent MP, while a mammoth swing in 2010 has given the Conservatives a fairly impregnable 10,000 majority in Dartford.

One bet could be Worcester (declaring 5am) which has a good record recently, going with the eventual winner at every election since 1979 and looks genuinely close this time. But it only saw its first Labour MP in 1997, and whether it has lost its Conservative roots, which otherwise date back to the 1920s, remains to be seen.

Another possibility is Gloucester (4am), which has gone with the winner at most elections since the 1960s. The required swing of 2.4% should be within Labour’s reach if it is going to win most MPs, although the party had a very poor showing here at last year’s local and European elections suggesting it could struggle to take it. Labour will hope that’s not the story for their night nationally too.


Turnout usually becomes clear earlier than results do. As my colleague Andy White pointed out on Tuesday, the Conservatives lead among all traditionally high-turnout groups, Labour among less high turnout demographics. A high turnout today should nominally then help Labour as it suggests more people who usually stay at home (and likely to be Labour leaning) have gone out and voted. The only caveat is that the last election to puncture the 70% threshold was 1992 when the Tories ended up convincing victors despite the polls suggesting otherwise – although the age profiles of the parties weren’t as polarised as they are now.

A really bad night for the Liberal Democrats is good news for the Conservatives

Lib Dem wipeout is a story many are expecting – however quite how bad their fate is will impact on the battle for largest party. The consensus seems to have them expected to lose around 30 seats - the first 27 of these should split evenly between nine Conservative, nine Labour and nine SNP. However, if the Lib Dems start losing seats beyond these, the vast majority would go to the Conservatives. Therefore, if Nick Clegg’s party lose a number of nominally safe, Tory-facing marginals early on (for example, Torbay 2am, Taunton Deane 2.30am, Eastbourne 3am), it would suggest it could be a long night for the party. It’s also worth mentioning that in the grand scheme of Coalition forming the aggregate number of Conservative and Liberal Democrat MPs is unchanged, giving no advantage to these parties' prospects of forming a Government together.

National vote shares - don’t confuse GB and UK vote shares

Important to remember: General Election results are reported on the parties’ UK share of the vote, but polls report the British shares of the vote only (i.e. excluding Northern Ireland). Their accuracy should therefore be measured against this rather than the UK figures. Which you are using might also be important to get straight in advance if you want to avoid arguments over the office sweepstake. In 2010, the Conservatives won 36% of the UK vote, against Labour’s 29%. The GB figures were 37% and 30%.

Labour/Conservative marginals are worth double

In the race to be largest party, it’s worth remembering that seats which switch between the main two parties are worth double – not only does it add one to your own tally, it also knocks one off your opponent’s.

Don’t confuse swing with size of majority

One of the easiest mistakes to make but one to get straight before election night: the swing required to win a seat is not the same as that seat’s majority - it’s half that figure. If the Conservatives lead Labour by ten points in a seat (let’s say 40% to 30%), Labour need five percent of the seat’s population who voted for the Conservatives at the last election to “swing” to Labour, taking the Conservatives down to 35% and at the same time taking Labour up to 35%. They then need one more vote than the Conservatives to win the seat, but this is usually left out of the swing figures as a single vote makes such a negligible difference in percentage terms.

So when Jeremy Vine’s talking about Labour needing a four point swing to win a seat, he’s referring to a seat where the incumbent leads Labour by 8%.

Why pay attention to the swingometer

The importance of swing was first identified in the 1940s by the godfather of British psephology, David Butler, when he realised that swing across each individual constituency tended to be remarkably consistent across the whole country.

Once upon a time, one could look at the results in the first seats to declare and calculate the swing from one party to another – say five points from Tories to Labour again. You could then look at how many seats the Tories held previously by less than ten points and would therefore fall to Labour this time on a five point swing too. Add this number to Labour’s previous seat tally, and take it away from the Tories’ and you had yourself a rough and ready final seat projection.

While this technique came a little unstuck during the Blair years when Labour would lose vast swathes of the voters in safe seats but hold onto marginal seats, Uniform National Swing performed well in 2010 (particularly if you split swing out nationally, and separated Scotland) and explains the enduring popularity of the swingometer.

Adding it all together

Tonight, we would advise not setting too much store by the very first constituencies to announce – the Sunderland seats have tended to swing slightly more violently than the nation as a whole (although the 5.75 point swing in Sunderland Central in 2010 was close to English national swing). Projecting what happened in Scotland may also be tricky, but if you average the swings from the first 10-20 English results to be announced, you may want to give the projection technique a go to estimate a rough number of Conservative marginals which Labour could pick up (or vice versa).

The table below shows how many Conservative seats would fall to Labour based on swings of different sizes. Polling seems to suggest a swing across Great Britain of 3.5 points, which would, in theory, see Labour pick up around 40 Conservative seats. That said, polling has also shown the swing in England only to be a little greater and approaching five points, which would suggest around 60 gains (that Labour are up around five points in vote share across Britain suggest a better performance in England and Wales to make up for the collapse in Scotland).

The table also includes what impact this would have on total seat tallies (by adding or subtracting the seat changes to each party’s current total). The third column shows what would happen before taking into account results in Liberal Democrat or Scottish seats, which shows Labour would only need a two point swing to win most seats (equivalent to 35% to 31% in terms of the national vote).

Of course, Scottish and Liberal Democrat constituencies are going to impact the results, so we’ve therefore included in the fourth column what could happen to the seat totals if the Liberal Democrats hypothetically lost 13 to the Conservatives, ten to Labour, and the SNP won 35 of Labour’s 40 Scottish seats. These are not projections and are included here for illustrative purposes only, but goes to show some way how tonight could pan out.

Swing Number of seats to switch Implied seat tally (before SNP gains / LD losses). Con seats / Lab seats
Implied seat with hypothetical SNP / LD resultsCon seats / Lab seats
1% 12 292 / 271 305 / 245
2% 25 279 / 284 292 / 258
3% 37 267 / 296 280 / 270
3.5% 40 264 / 297 277 / 273
4% 47 257 / 306 270 / 280
5% 60 244 / 319 257 / 293

Generally, the key tally to keep track of in the race for largest party is if Labour can win around at least 45 Conservative seats and lose no more than 35 to the SNP. If they miss either of these targets, they will need to make it up on the other. Similarly if the Tories completely wipe out the Liberal Democrats and Nick Clegg’s party looks as if they are going to win less than a score of seats, Labour will also have to over-perform on both to make this up.

What are the seats declaring early that might give us an indication of whether Labour can overturn the Conservatives as the largest party? While keeping one eye on Scotland, a good bet might be those requiring the 3-4% swing Labour needs to take them to 45 Tory gains: Pendle (01.30), South Swindon (02.30) Wirral West (02.30), Cannock Chase (03.00), Stevenage (03.30) and Harrow East (04.00). Barring a miraculous recovery in Scotland, if Labour miss most of these, Cameron will most likely be making the first move tomorrow morning.