Exit polling in the UK is difficult in any case because there is no official information on the votes cast at individual polling stations, there is no consistent relationship between votes and seats, postal voting is common, and one in six people refuse to answer pollsters’ questions.
This year the difficulties will be greater because the 2005 sample of polling districts is short of Labour-Liberal Democrat marginal seats, the 2001 census is out of date and boundaries have been changed. It means that the boffins have a good chance of successfully estimating the loss of Labour seats to the Conservatives but they will find it harder to evaluate any Liberal Democrat surge.
I have just frigged around with my swingie-thing and found I had to put a big negative percentage in the LibDem box, obviously. If you follow what Chris Giles says and keep the Conservative – Labour gains the same, but boost so the Lib Dems get 26 instead of 23%, you get 79 seats for the LibDems, gained off Labour