Track Record

Backtested against the 2024 General Election. Each model improvement narrows the error. Transparency about what we get wrong builds trust in what we get right.

Improvement journey

Phase 1

Basic UNS model

84.9%
552/650 seats correct

Uniform national swing from polling averages. Wrong data for some seats, no house effect correction.

Lab -45Con +38Reform -12LD +8Green -3SNP +14
Phase 1.5

House effects + calibration

98.2%
638/650 seats correct

Added house effect correction, poll recency weighting, and regional calibration. All 12 remaining errors were tight three-way marginals.

Lab -5Con +3Reform -2LD +2SNP +2
Phase 2a

Full model pipeline

99.4%
646/650 seats correct

Vote source models, turnout adjustment, tactical voting. The 4 remaining errors are all independent wins that no national model can predict.

Lab -1Con +1Green -1SNP +1

The 4 we still get wrong

All four errors are independent wins driven by hyperlocal dynamics invisible to any national polling model. This is the honest limit of what polling-based forecasting can achieve.

Leicester South
Actual: IndependentPredicted: Lab

Hyperlocal independent candidacy with strong community support

Dewsbury & Batley
Actual: IndependentPredicted: Lab

Independent candidate with local profile beyond polling capture

Birmingham Perry Barr
Actual: IndependentPredicted: Lab

Local independent campaign invisible to national model

Blackburn
Actual: IndependentPredicted: Lab

Strong independent challenger not captured by any polling

What the backtest tells us

The model correctly assigns seats in 99.4% of cases using only pre-election polling data. The multi-layer approach (regional swing + vote source + turnout + tactical) adds 94 correct seats over simple uniform national swing.

The remaining 4 errors represent the genuine floor of national-model accuracy. Independent candidates are a local phenomenon that no amount of polling sophistication can predict. We show this honestly rather than claiming perfect accuracy.