ABSTRACT
The next election will be substantively fought and won in a relatively small number of 'marginal' seats. The traditional concept of a marginal seat is not particularly helpful in studying the electoral landscape of today, as it is based on a rough rule of thumb established in the days when electoral competition in Britain was dominated by just two parties. Then, if the gap between the two parties' votes was less than 10 per cent of the total vote, the seat was considered to be vulnerable to change. The constituency results for every general election since 1955 have been examined here to determine more precisely, using probabilities based on previous general elections, the gap that is required for a seat to become vulnerable. In many cases it is found that the size of that gap is related to the performance of the third party.
