GDP, quarterly
2026Q380% band +0.23 … +0.88
Backtest: MAE 0.23 · 69% of AR error · direction 87% · n=69

ONS faster indicators, Bank of England, HMRC, energy, traffic, and price data
Model estimates of not-yet-published official statistics, updated daily through each release cycle. Accuracy shown is each model’s own out-of-sample backtest.
80% band +0.23 … +0.88
Backtest: MAE 0.23 · 69% of AR error · direction 87% · n=69
80% band -0.59 … +0.44
Backtest: MAE 0.326 · 96% of AR error · direction 62% · n=207
80% band -0.62 … +2.05
Backtest: MAE 0.977 · 101% of AR error · direction 53% · n=40
80% band -0.13 … +0.10
Backtest: MAE 0.076 · 94% of AR error · direction 37% · n=172
Weekly monitoring index (52-week transforms). Buys timeliness and turning-point speed, not calm-period precision; prone to false signals and to base-effect echoes a year after large shocks.
Staggered combination of an official-data block, a survey/indicator bridge and an autoregressive benchmark, re-weighted through the data release cycle. Backtests are walk-forward against the latest vintage; pandemic-window observations are excluded from estimation and shown separately. Sources: ONS, Bank of England, public high-frequency data.
High-speed series (gold, left axis) shown against the official statistics they lead (dark, right axis). Last 4 years.
NSA & SA — ONS / EUROCONTROL
Fuel volume index — ONS
Ship visits at major UK ports — ONS / exactEarth
Monthly car + LCV registrations by sale type — ONS / SMMT
NSA index, 2023 = 100 — ONS / BT Active Intelligence
7-day rolling average, NSA, index 2023 = 100 — ONS / Revolut
Monthly DD failure rate, SA — ONS / Pay.UK