*UPDATE* Since the latest polls showing the Lib Dems over 30%, this post has dated considerably since I wrote it.
In the 1997 general election, 180 seats changed hand. This was by a big margin a post-war record, helped by boundary changes agreed by the Tories in the nineties that propped up their safest seats but helped Labour collect a number of marginals they wouldn't otherwise have won.
This was an inept 'me first' strategy on constituency boundaries from Tory grandees protecting themselves at the cost to the overall party. The Tories changed this to a more 'socialist' strategy in the reviews before this election and this will help them regain some seats.
In total since 1992, the Tories lost 150 seats to Labour and 30 to the Lib Dems. In 2005 they reclaimed 30 seats back from Labour and 5 from the Lib Dems. So, this leaves around 120 Labour held marginals and 25 Lib Dem marginals that are vulnerable and that the Tories need to win nearly all of these back to get a working majority. These seats effectively are the only ones that matter in this election. Indeed the pattern is that over many decades - only 250 seats (out of 650) will EVER change. Effectively voters in 75-80% of seats are irrelevant to the result.
There are a number of patterns to these vulnerable seats. Psephologists refer to most of them as the METTHs - 'Medium English Towns and Their Hinterlands' - the Bolton and Lutons, Burton and Hastings, Brighton and Crewes etc.
Another noticeable factor is the number of votes required to win these seats. Labour won these seats with nigh on 30,000 votes each seat in 1997, since then their winning vote has dropped to about half that, to around 15,000-18.000. The Tory vote has been remarkably consistent, stuck at around 13,000-17,000 since 1997. The polls suggest this vote will rise by little if anything in 2010. Also demographics suggest this to be the case - the Tory vote is largest amongst the over 55s and 3m of these voters have died since 2005. Of the 4 million new voters, the Tories are not attracting enough to compensate. Despite the Ashcroft millions and Tory press and hence the 'never voted Tory' campaign.
So, the bottom line is that - if the Labour vote drops near to or below 13,000, these seats are almost certainly lost. Remember all the seats average around 68,000 electors - there is little variation from this - despite the Tory squeals about the anomaly of the Western Isles and the Isle of Wight - these are unusual seats and only a handful of seats. Of the 68,000 electors about 38,000 still bother to vote in these tightly fought marginals. Lets look at some examples:-
Bolton North East
Hastings & Rye
Obviously interesting things are happening in Brighton Pavilion with the Greens and local factors will make a difference to results that the national press/ pollsters won't pick up on. I suggest any Lab/Tory marginal where Labour enjoy less than a 3,000 majority will be lost - this means the Tories could gain 80-90 or so of the 120 seats, with maybe 5-10 gains from the Lib Dems. The Tory/LIb Dem fight is going to be harder to predict but I think that the Libs could hold most of these (especially since Nick Clegg seems to be doing well in the TV debates).
The polls are pointing towards national vote shares of CON 37%, LAB 31% and LIB 21% (although this could change after Clegg's performance). Ironically of course a vote for the Lib Dems in most of the 150 seats that matter might actually reduce chances of a hung parliament and increase a Tory win.
Anyway, using these poll figures, the direction of travel of Tory and Labour votes in marginals since 1997 (Tory vote static, Labour losing 20% per election), demographics, local council results and some guesswork I predict the following:-
Turnout will remain the same as 2005 - about 27m will vote out of a possible 45m. 2005 figures in brackets:-
Tories 10m (8.8m) votes - 310 seats
Labour 8.4m (9.6m) votes - 240 seats
LibDem 5.7m (6.0m) votes - 60 seats
UKIP 1.1m (0.6m) votes - 0 seats
Green 0.5m (0.3m) - 1 seat
Others 1.8m (1.8m)- 39 seats
This will leave the Tories 20 or so short of a majority.
So, 13,000 votes is the magic number in a marginal, it seems 10m votes (or just 22% of electorate) is the magic number to get close to a majority of seats (Labour managed it with 9.6m in 2005)
As we can see from the above seat numbers - any talk of a Lib/Lab government is fantasy as they will not have enough seats between them to form a government. Only a Lib/Tory government or a Tory government is possible on current poll numbers.
Locally I predict the following:-