Election Night Running Totals, 1979—2010

You should all have @AndyJSajs in your list of friends on Twitter

This is his fantastic work.

Click the years below for a document with the running totals.

2010
2005
2001
1997
1992
1987
1983
1979

(Some of the late results in 1979 are a bit uncertain because they weren't all flashed on the screen, particularly Yarmouth and Thurrock which I don't think actually were the final two results in that election. If anyone has any further information on the 1979 declaration times I'd be grateful to know about it).


The interesting question IMO is how many results are needed to give the eventual national swing to a particular level of accuracy.

Useful measures might be (a) within 2%, (b) within 1%, (c) within 0.5%.

Results:

2010:
(a) 16
(b) 26
(c) 135

2005:
(a) 1
(b) 4
(c) 91

2001:
(a) 1
(b) 8
(c) 89

1997:
(a) 1
(b) 99
(c) 341

1992:
(a) 1
(b) 16
(c) 18

1987:
(a) 5
(b) 7
(c) 76

1983:
(a) 1
(b) 1
(c) 11

1979:
(a) 2
(b) 2
(c) 372


Summary: the greatest number of results needed to give the eventual national swing within 2% was 16 in 2010, within 1% was 99 in 1997, and within 0.5% was 372 in 1979. 1983 was the best election for getting the quickest indication of the eventual national swing with just 11 results needed to give it within 0.5% (and 1 result for 1% and 2%).