by Dr Adrian Worton
Even before the 2019 General Election was called, there have been calls for nationwide tactical voting to deny the Conservatives a majority. This has been brought into focus by polling suggesting that the Conservatives are indeed on course for a majority, but one that could be chipped away by tactical voting in key constituencies.
There have been many attempts at creating tactical voting guides. Most recently, the Observer published its tactical voting guide, focusing on 50 seats. As with all guides, there have been complaints over the recommendations. We are going to look through these 50 seats today and assess how accurate the guide is.
Even before the 2019 General Election was called, there have been calls for nationwide tactical voting to deny the Conservatives a majority. This has been brought into focus by polling suggesting that the Conservatives are indeed on course for a majority, but one that could be chipped away by tactical voting in key constituencies.
There have been many attempts at creating tactical voting guides. Most recently, the Observer published its tactical voting guide, focusing on 50 seats. As with all guides, there have been complaints over the recommendations. We are going to look through these 50 seats today and assess how accurate the guide is.
Who to trust?
The problem is, what do we use to measure whether a recommendation is good or bad? Everyone will have received various election leaflets from parties standing in their seat, and these leaflets can be particularly confusing. For example, these are snippets of leaflets I have received, (my seat is Dunfermline & West Fife):
The problem is, what do we use to measure whether a recommendation is good or bad? Everyone will have received various election leaflets from parties standing in their seat, and these leaflets can be particularly confusing. For example, these are snippets of leaflets I have received, (my seat is Dunfermline & West Fife):
Here, all three of the Conservatives, Labour and the Liberal Democrats are claiming they are the only party who can challenge the SNP incumbent. In particular, the Conservatives and Lib Dems are using percentage change to boost how they look in the chart. This is a horrendous use of statistics - basically, it's easier to boost your vote from a lower base.

We are going to use our own model's probabilities (taken from bookmaker odds), but our main source will be the comprehensive YouGov MRP model. This model was the best estimator of the previous election, and provides vote ranges for each seat.
For example, their estimates for Dunfermline & West Fife are shown opposite. This suggests that this is a relatively safe SNP seat, but it's not impossible that the Conservatives or Labour could take it. It particularly emphasises how dishonest the Lib Dems' leaflet was, in keeping with their election communications elsewhere.
It should also be noted that YouGov last updated their MRP on the 27th November; since then there have been changes in the polls. Using Britain Elects' poll tracker, the Conservatives have stayed static, Labour have gained around 2.5% whilst the Lib Dems have lost 1.3%. So in cases where the MRP is close between Labour and the Lib Dems, it would be sensible to err on the side of Labour.
We will also occasionally refer to tactical voting advice by Best for Britain and tactical.vote.
So, let's go through the Observers' recommendations. The recommendations are made from the point of view of "anyone but Tory", which is how we'll rate them.
Obvious choices
In 34 seats, the choice is obvious, as the seat is a two-horse race between the Conservatives and someone else. These are (grouped by who to vote for):
Labour: Chingford & Woodford Green, Chipping Barnet, Filton & Bradley Stoke, Hendon, Loughborough, Portsmouth South, Truro & Falmouth, Uxbridge & South Ruislip, Watford, Wycombe
Liberal Democrats: Cheadle, Cheltenham, Esher & Walton, Guildford, Hazel Grove, Lewes, Richmond Park, South Cambridgeshire, St Albans, St Ives, Totnes, Wantage, Winchester, Wokingham
SNP: Aberdeen South, Angus, Ayr Carrick & Cumnock, Banff & Buchan, Dumfries & Galloway, East Renfrewshire, Gordon, Moray, Ochil & South Perthshire, Stirling
These seats have a simple equation: by voting against the recommendation it makes it far more likely the Conservatives will win the seat. Of course, voters may be uncomfortable voting for particular non-Conservative parties (particularly in Scotland, where the independence divide is strong, and where the SNP will claim all votes for them as part of their mandate for a new referendum), but this doesn't change the reality of their seat's arithmetic. Now to the remaining seats:
Beaconsfield
For example, their estimates for Dunfermline & West Fife are shown opposite. This suggests that this is a relatively safe SNP seat, but it's not impossible that the Conservatives or Labour could take it. It particularly emphasises how dishonest the Lib Dems' leaflet was, in keeping with their election communications elsewhere.
It should also be noted that YouGov last updated their MRP on the 27th November; since then there have been changes in the polls. Using Britain Elects' poll tracker, the Conservatives have stayed static, Labour have gained around 2.5% whilst the Lib Dems have lost 1.3%. So in cases where the MRP is close between Labour and the Lib Dems, it would be sensible to err on the side of Labour.
We will also occasionally refer to tactical voting advice by Best for Britain and tactical.vote.
So, let's go through the Observers' recommendations. The recommendations are made from the point of view of "anyone but Tory", which is how we'll rate them.
Obvious choices
In 34 seats, the choice is obvious, as the seat is a two-horse race between the Conservatives and someone else. These are (grouped by who to vote for):
Labour: Chingford & Woodford Green, Chipping Barnet, Filton & Bradley Stoke, Hendon, Loughborough, Portsmouth South, Truro & Falmouth, Uxbridge & South Ruislip, Watford, Wycombe
Liberal Democrats: Cheadle, Cheltenham, Esher & Walton, Guildford, Hazel Grove, Lewes, Richmond Park, South Cambridgeshire, St Albans, St Ives, Totnes, Wantage, Winchester, Wokingham
SNP: Aberdeen South, Angus, Ayr Carrick & Cumnock, Banff & Buchan, Dumfries & Galloway, East Renfrewshire, Gordon, Moray, Ochil & South Perthshire, Stirling
These seats have a simple equation: by voting against the recommendation it makes it far more likely the Conservatives will win the seat. Of course, voters may be uncomfortable voting for particular non-Conservative parties (particularly in Scotland, where the independence divide is strong, and where the SNP will claim all votes for them as part of their mandate for a new referendum), but this doesn't change the reality of their seat's arithmetic. Now to the remaining seats:
Beaconsfield

Observer recommendation: Other (Dominic Grieve)
TGIAF probabilities: CON 90%, IND 10%
This is a strange seat for the Observer to pick. Clearly, Grieve is the only one who can deny the Conservatives here, but realistically it appears as though there is little chance of him doing so.
There are a lot of seats that are a lot closer, so this feels like it would have been more sensible to feature a different recommendation.
Chelsea & Fulham
TGIAF probabilities: CON 90%, IND 10%
This is a strange seat for the Observer to pick. Clearly, Grieve is the only one who can deny the Conservatives here, but realistically it appears as though there is little chance of him doing so.
There are a lot of seats that are a lot closer, so this feels like it would have been more sensible to feature a different recommendation.
Chelsea & Fulham

Observer: Lib Dem
TGIAF: CON 94%, LIB 4%, LAB 2%
It's hard to disagree with the Observer here, as there is a minor chance of the Lib Dems challenging the Conservatives here.
Crucially, there is a healthy chunk of Lavour votes that, if they were to switch to the Lib Dems, would drastically alter the complexion of this race.
Cities of London & Westminster
TGIAF: CON 94%, LIB 4%, LAB 2%
It's hard to disagree with the Observer here, as there is a minor chance of the Lib Dems challenging the Conservatives here.
Crucially, there is a healthy chunk of Lavour votes that, if they were to switch to the Lib Dems, would drastically alter the complexion of this race.
Cities of London & Westminster

Observer: Lib Dem
TGIAF: CON 77%, LIB 21%, LAB 2%
YouGov slightly favour Labour here, and that is only strengthened when you factor in the recent changes in polling. Labour also finished 10,000 votes higher than the Lib Dems last time.
This does make you wonder whether the Observer backs the Lib Dems because the Lib Dem candidate is Chuka Umunna, a prominent Remain figure. This said, their own polling saw the Lib Dems in second, although this also pre-dates Labours’ recent poll improvement.
It’s worth noting that tactical.vote recommend Labour, and whilst Best for Britain recommend the Lib Dems, their most recent projection for the seat sees Labour as 2nd-placed. Overall, and despite our own model's probabilities, Labour seem the best bet here.
East Lothian
TGIAF: CON 77%, LIB 21%, LAB 2%
YouGov slightly favour Labour here, and that is only strengthened when you factor in the recent changes in polling. Labour also finished 10,000 votes higher than the Lib Dems last time.
This does make you wonder whether the Observer backs the Lib Dems because the Lib Dem candidate is Chuka Umunna, a prominent Remain figure. This said, their own polling saw the Lib Dems in second, although this also pre-dates Labours’ recent poll improvement.
It’s worth noting that tactical.vote recommend Labour, and whilst Best for Britain recommend the Lib Dems, their most recent projection for the seat sees Labour as 2nd-placed. Overall, and despite our own model's probabilities, Labour seem the best bet here.
East Lothian

Observer: SNP
TGIAF: SNP 71%, LAB 15%, CON 14%
The Conservatives are third-favourites in this seat, so it’s a slightly odd choice to focus on for a case of tactical voting. It’s hard to tell either the SNP or Labour to stand aside in a seat they have a good chance of winning (and where Labour are the incumbents).
The only tactical advice here is to suggest Lib Dem voters back either the SNP or Labour. It’s unfair to ask SNP or Labour voters to switch.
Finchley & Golders Green
TGIAF: SNP 71%, LAB 15%, CON 14%
The Conservatives are third-favourites in this seat, so it’s a slightly odd choice to focus on for a case of tactical voting. It’s hard to tell either the SNP or Labour to stand aside in a seat they have a good chance of winning (and where Labour are the incumbents).
The only tactical advice here is to suggest Lib Dem voters back either the SNP or Labour. It’s unfair to ask SNP or Labour voters to switch.
Finchley & Golders Green

Observer: LIB
TGIAF: CON 87%, LIB 11%
Finchley & Golders Green is a noteworthy seat, as Jewish former Labour MP Luciana Berger stands for the Lib Dems in a seat with a very high Jewish population. With Labour’s antisemitism problems, it is understandable why the Observer (and Best for Britain) back the Lib Dems.
However, the Lib Dems took under 7% at the last election, and they would require the largest swing in election history to take the seat. Ultimately, it is hard to see anything other than a Conservative win here.
Kensington
TGIAF: CON 87%, LIB 11%
Finchley & Golders Green is a noteworthy seat, as Jewish former Labour MP Luciana Berger stands for the Lib Dems in a seat with a very high Jewish population. With Labour’s antisemitism problems, it is understandable why the Observer (and Best for Britain) back the Lib Dems.
However, the Lib Dems took under 7% at the last election, and they would require the largest swing in election history to take the seat. Ultimately, it is hard to see anything other than a Conservative win here.
Kensington

Observer: LIB
TGIAF: CON 69%, LAB 22%, LIB 10%
Whilst YouGov’s MRP has a slender Lib Dem lead over Labour, this is likely to have been eliminated by recent poll swings. Labour took this seat with a very narrow margin in 2017. The Observer’s recommendation disagrees with tactical.vote, who say Labour, whilst Best for Britain haven’t made a call.
Crucially here, it will be difficult to ask Labour voters to vote for Lib Dem candidate Sam Gyimah, who is a former Conservative minister who has made some truly atrocious comments regarding the Labour incumbent Emma Dent Coad and the Grenfell disaster. It feels more likely that Lib Dem voters could be better convinced to switch from Gyimah, who was part of Theresa May’s pro-Brexit cabinet, to Dent Coad who has rebelled against the Labour whip in favour of pro-Remain positions.
Ultimately, therefore, we would recommend Labour. However, it’s now likely that the conflicting advice will make this a comfortable Conservative gain.
Kirkcaldy & Cowdenbeath
TGIAF: CON 69%, LAB 22%, LIB 10%
Whilst YouGov’s MRP has a slender Lib Dem lead over Labour, this is likely to have been eliminated by recent poll swings. Labour took this seat with a very narrow margin in 2017. The Observer’s recommendation disagrees with tactical.vote, who say Labour, whilst Best for Britain haven’t made a call.
Crucially here, it will be difficult to ask Labour voters to vote for Lib Dem candidate Sam Gyimah, who is a former Conservative minister who has made some truly atrocious comments regarding the Labour incumbent Emma Dent Coad and the Grenfell disaster. It feels more likely that Lib Dem voters could be better convinced to switch from Gyimah, who was part of Theresa May’s pro-Brexit cabinet, to Dent Coad who has rebelled against the Labour whip in favour of pro-Remain positions.
Ultimately, therefore, we would recommend Labour. However, it’s now likely that the conflicting advice will make this a comfortable Conservative gain.
Kirkcaldy & Cowdenbeath

Observer: LAB (initially SNP)
TGIAF: LAB 70%, IND 28%, CON 2%
This is another unusual seat, as the SNP have withdrawn their support for their candidate following allegations of antisemitism and transphobia (he will, however, remain on the ballot). The YouGov polling opposite pre-dates this change, and Observer recommendation initially backed the SNP before correcting to Labour in light of this.
With this in mind, Labour are the obvious tactical vote here.
Midlothian
TGIAF: LAB 70%, IND 28%, CON 2%
This is another unusual seat, as the SNP have withdrawn their support for their candidate following allegations of antisemitism and transphobia (he will, however, remain on the ballot). The YouGov polling opposite pre-dates this change, and Observer recommendation initially backed the SNP before correcting to Labour in light of this.
With this in mind, Labour are the obvious tactical vote here.
Midlothian

Observer: SNP
TGIAF: SNP 88%, LAB 10%, CON 2%
Like its neighbour East Lothian, this is another seat where it is unclear why this has been selected for tactical voting advice, as the Conservatives sit in third position.
Again, the only advice would be for Lib Dem voters to switch to either the SNP or Labour.
Putney
TGIAF: SNP 88%, LAB 10%, CON 2%
Like its neighbour East Lothian, this is another seat where it is unclear why this has been selected for tactical voting advice, as the Conservatives sit in third position.
Again, the only advice would be for Lib Dem voters to switch to either the SNP or Labour.
Putney

Observer: LAB
TGIAF: CON 58%, LAB 40%, LIB 2%
This seat was held by prominent Conservative minister Justine Greening until she was kicked out of the party. Unlike other Conservative rebels, Greening isn’t contesting her old seat. Labour were a close second last time, and look best-placed to challenge again.
There are a lot of Lib Dem votes remaining here, and the Observer advice for them to switch to Labour is correct, as this is a very winnable seat for Labour.
Rushcliffe
TGIAF: CON 58%, LAB 40%, LIB 2%
This seat was held by prominent Conservative minister Justine Greening until she was kicked out of the party. Unlike other Conservative rebels, Greening isn’t contesting her old seat. Labour were a close second last time, and look best-placed to challenge again.
There are a lot of Lib Dem votes remaining here, and the Observer advice for them to switch to Labour is correct, as this is a very winnable seat for Labour.
Rushcliffe

Observer: LAB
TGIAF: CON 96%, LIB 2%, LAB 2%
As with Putney, this seat’s MP was a prominent Conservative former minister, this time Ken Clarke, until he lost the whip.
The Observer is correct that Labour is the tactical vote here, but this is a seat where a non-Conservative victory seems unlikely, making this another strange choice to be picked out as part of their recommendations.
Sheffield Hallam
TGIAF: CON 96%, LIB 2%, LAB 2%
As with Putney, this seat’s MP was a prominent Conservative former minister, this time Ken Clarke, until he lost the whip.
The Observer is correct that Labour is the tactical vote here, but this is a seat where a non-Conservative victory seems unlikely, making this another strange choice to be picked out as part of their recommendations.
Sheffield Hallam

Observer: LIB
TGIAF: LIB 83%, LAB 16%
As with the two Lothian seats, this is one where the Conservatives are almost certainly not winning (1% in our model).
Therefore, it is misleading of the Observer to highlight this as a key tactical voting seat. Neither Lib Dem nor Labour voters should be talked into voting for the other to keep the Conservatives out.
South West Hertfordshire
TGIAF: LIB 83%, LAB 16%
As with the two Lothian seats, this is one where the Conservatives are almost certainly not winning (1% in our model).
Therefore, it is misleading of the Observer to highlight this as a key tactical voting seat. Neither Lib Dem nor Labour voters should be talked into voting for the other to keep the Conservatives out.
South West Hertfordshire

Observer: IND
TGIAF: CON 96%, IND 3%
This is a virtually-identical seat to Beaconsfield, as another former Conservative minister, this time David Gauke, stands to retain his seat.
We’ve highlighted this one because it’s yet another one where you wonder why the Observer bothered to make a recommendation.
Southport
TGIAF: CON 96%, IND 3%
This is a virtually-identical seat to Beaconsfield, as another former Conservative minister, this time David Gauke, stands to retain his seat.
We’ve highlighted this one because it’s yet another one where you wonder why the Observer bothered to make a recommendation.
Southport

Observer: LAB
TGIAF: CON 88%, LAB 8%, LIB 4%
With recent poll changes, Labour probably have more of a chance here than the YouGov data opposite suggests, particularly if Lib Dem voters can be minded to switch.
The Observer is right to recommend Labour here.
Wimbledon
TGIAF: CON 88%, LAB 8%, LIB 4%
With recent poll changes, Labour probably have more of a chance here than the YouGov data opposite suggests, particularly if Lib Dem voters can be minded to switch.
The Observer is right to recommend Labour here.
Wimbledon

Observer: LIB
TGIAF: CON 78%, LIB 21%
Our model doesn’t rate Labour’s chances at all in this seat, although if you apply recent polling changes to this seat, they are probably close to the Lib Dems’ level.
However, it’s understandable that the Observer has recommended the Lib Dems. Best for Britain has also backed the Lib Dems, whilst tactical.vote backs Labour. It feels like mixed messages will result in a Conservative hold.
Ynys Mon
TGIAF: CON 78%, LIB 21%
Our model doesn’t rate Labour’s chances at all in this seat, although if you apply recent polling changes to this seat, they are probably close to the Lib Dems’ level.
However, it’s understandable that the Observer has recommended the Lib Dems. Best for Britain has also backed the Lib Dems, whilst tactical.vote backs Labour. It feels like mixed messages will result in a Conservative hold.
Ynys Mon

Observer: LAB
TGIAF: PLA 38%, CON 34%, LAB 28%
This is another nasty one where both Labour and Plaid Cymru would claim to be the challengers. Both our model and YouGov’s data suggest this is a very close three-way marginal.
This time, there is a tactical voting consensus, as tactical.vote and Best for Britain both back Labour. Therefore, in a slightly tautological manner, Labour should be the tactical vote here.
York Outer
TGIAF: PLA 38%, CON 34%, LAB 28%
This is another nasty one where both Labour and Plaid Cymru would claim to be the challengers. Both our model and YouGov’s data suggest this is a very close three-way marginal.
This time, there is a tactical voting consensus, as tactical.vote and Best for Britain both back Labour. Therefore, in a slightly tautological manner, Labour should be the tactical vote here.
York Outer

Observer: LAB
TGIAF: CON 93%, LIB 3%, LAB 3%
The Observer’s rationale for including this safe Conservative seat in its list of recommendations is that the student vote could tip things.
It’s unclear if there’s strong evidence that there is a major student group living here (one would have thought most students would be in the most urban York Central constituency), but either way this appears another odd constituency to bother with
Conclusion
There’s no strong evidence that the Observer has major biases in its recommendations, although there are a few cases where they have backed the Lib Dems when Labour appear a more credible option. There are also a few Labour-held seats that the Observer has picked to recommend a challenger, despite Labour remaining competitive.
What is striking is the number of constituencies picked out where it appears a comfortable Conservative hold. It’s clearly an editorial decision to have exactly 50 seats, so these seats will have been picked at the expense of others.
There are plenty of seats this election with everything to play for. Many of these are likely to be much closer than some of the seats chosen by the Observer. To fail to give tactical voting recommendations here is a poor decision. Therefore, we will go through a list of our own this week, continuing to draw on the same sources.
TGIAF: CON 93%, LIB 3%, LAB 3%
The Observer’s rationale for including this safe Conservative seat in its list of recommendations is that the student vote could tip things.
It’s unclear if there’s strong evidence that there is a major student group living here (one would have thought most students would be in the most urban York Central constituency), but either way this appears another odd constituency to bother with
Conclusion
There’s no strong evidence that the Observer has major biases in its recommendations, although there are a few cases where they have backed the Lib Dems when Labour appear a more credible option. There are also a few Labour-held seats that the Observer has picked to recommend a challenger, despite Labour remaining competitive.
What is striking is the number of constituencies picked out where it appears a comfortable Conservative hold. It’s clearly an editorial decision to have exactly 50 seats, so these seats will have been picked at the expense of others.
There are plenty of seats this election with everything to play for. Many of these are likely to be much closer than some of the seats chosen by the Observer. To fail to give tactical voting recommendations here is a poor decision. Therefore, we will go through a list of our own this week, continuing to draw on the same sources.