Old model

We’ll continue to publish weekly updates using our old forecast model, for comparison and consistency.

Forecast b

Date of forecast: 13 March 2015
Days till the election: 55

Inputted current average poll shares
Con: 33%
Lab: 33%
LD: 8%
Others (inc. UKIP): 26%
– UKIP: 15%

Forecast Election Day Shares (with 95% Prediction Intervals)
Con: 34.4% (±5.8, i.e. 29% – 40%)
Lab: 31.5% (±4.4, i.e. 27% – 36%)
LD: 10.0% (±5.1, i.e. 5% – 15%)
Implied point estimate shares for:
– Others (inc. UKIP): 24.1%
– UKIP: 13.0%

Forecast Election Day Seats (with approximate 95% Prediction Intervals)
Con: 296 (233 – 367)
Lab: 298 (230 – 358)
LD: 25 (21 – 31)
(Prediction intervals assume LD & others shares at central forecast, Con & Lab shares vary as per prediction intervals)

Central forecast: Lab largest party, but short of a majority by 28

Approximate probabilities of key outcomes
Con largest: 49%
… with a majority: 19%
Lab largest: 51%
… with a majority: 20%
Hung Parliament: 61%
… with Con largest: 30%
… with Lab largest: 31%
(probabilities may not sum due to rounding)

Shares trend

Seats trend

Probabilities trend

Probabilities trend b

Shares trend b

Seats trend b

One thought on “Old model”

  1. What happened to weekly updates with the old model? (which might b more accurate not including this massive incumbancy swing)

Leave a comment

Election analysis and forecasting