Sir Robert Worcester is a Visiting Professor of Government at LSE and an Honorary Fellow. He founded MORI in 1969.

The following post first appeared in the Observer on 2 May.

A year ago the Conservatives enjoyed a 14-point lead over Labour, 43% to 29%. The Liberal Democrats were a poor third, at 18%. Today, as they have been since the first TV leaders’ debate, the Conservatives are down eight points, the Lib Dems are up nine, and Labour down just one, a swing of 8.5% from the Conservatives to the Lib Dems with Labour hardly touched.

The bad news for David Cameron is that he was the future once. The bad news for Gordon Brown is that now, instead of his party running second, they are in third place in some recent polls and are fighting for their future.

One or other of these two men will be in No 10 next week, most likely after giving Nick Clegg whatever he wants in return for his support. Yet Clegg’s party will have run a bad third in seats won, just over 100 if he’s lucky, 65-70 if not.

This likely outcome is the consequence of a voting system that in the past two elections has rewarded the Lib Dems’ 23% share of the vote in 2005 with just 10% of the seats in the House of Commons, and their 19% share of the 2001 popular vote with only 8% of MPs.

The Observer and Ipsos MORI have therefore joined forces to provide voters with information they have not had in past elections. This is guidance on how they might cast their ballots tactically to get the outcome they want, whether that is to strengthen their party’s support nationally, or to ensure the candidate they wish to win does win, or to keep someone they don’t want to win from winning.

A tactical vote might be used by a Labour supporter in a Tory/Lib Dem marginal to boost Lib Dem momentum and oust the sitting Tory. Or, in another seat, for example, a Labour/Tory marginal where the Lib Dems are no-hopers, a Lib Dem supporter might vote to keep Tories from ousting a sitting Labour member if that voter hoped their party would support Labour in a hung parliament. There are obviously a number of permutations.

Over the past month Reuters has commissioned Ipsos MORI to carry out four surveys in Labour/Conservative marginals, so-called “battleground constituencies” where Labour-held seats would fall to the Conservatives with a swing of between 5% and 9%. Nick Clegg’s surprise runaway victory at the first leaders’ debate changed all this, however. Now, there is a strong possibility not of a Tory gain in these seats, but a win for the Lib Dems. The best bets are shown in the first table to the right, highlighted in yellow – 28 seats in all.

But few of the people who live in a marginal constituency of this kind know what power they could wield, maybe as few as 30%. The challenge to the political parties is to let their supporters in these seats know by next Thursday how important their votes could be. For example, in these marginals eight voters in 10 say that, even if they thought their candidate did not stand much chance of winning in their constituency, they would still stick with their party, and only one person in seven said they would switch their vote to another party.

In most elections, the outcome is determined by about four voters in a hundred. They are the “floating voters”. Around the country, one in five voters is a floating voter – but among that group it is only the one in five of those who live in the marginal constituencies whose votes count.

tables may include the constituency where you live. If it is there, then you are in one of the seats where your vote can have real power. And if you are willing to use your vote tactically, you may find yourself that one voter in 25 who will determine which party comes first, second or third, and even which parties will make up the uneasy alliance that now looks the most likely outcome on Thursday – a “hung” parliament.

But bear in mind that these tables are based not on what will happen on election day, but on the standing of the parties since the third debate on 29 April. Who knows what may happen in these last few days?

After all, my first election in Britain was 1970, lost in the last week by a reversal in the balance of payments, to the surprise of prime minister Harold Wilson. In fact, around 10% of voters in elections going back decades have told us that they made their final decision on whom they would be voting for in the final 24 hours of the election.

Turnout is expected to be sharply increased in Thursday’s election. That must be good for democracy, whatever the outcome.

If Thursday’s election replicates the average of all the polls taken since the third debate, the Conservatives would be about 40 seats short of the 326 needed to give them a majority in the House of Commons, while the Labour party would be short nearly 80.

Even so, if this were the outcome, it would be Brown’s constitutional right to try to form a government which would have sufficient support. “Sufficient” would mean that he was able to carry a majority in the house for his budget or the Queen’s speech and withstand a challenge from the opposition.

If he were to decide that he could not, then the leader of the party with the most seats, in this example David Cameron, would be asked to form a government. In practice, it is more likely that the leaders would spend a frantic weekend, as they did in early March 1974, trying to cobble together some sort of deal so that by Monday there would have been a visit to the Buckingham Palace and a government in place.

If this were done by either party with the support of the Lib Dems, then a new government would be able to face an opposition with a combined majority of at least a score of MPs with Labour – Lib Dem activists’ preferred partner – or even 80 with the Tories.

Sir Robert Worcester is the founder of MORI