Public forecast 004 · permanent record

Will the ONS headline UK unemployment rate published on 18 August 2026 be between 4.6% and 5.2% inclusive?

RESOLVED — scored against the source named before the outcome was known.

Question
Will the ONS headline UK unemployment rate published on 18 August 2026 be between 4.6% and 5.2% inclusive?
Probability band
86–94% — issued 2026-08-14, data cut-off 2026-08-14
Frozen criterion
Resolves TRUE if the headline unemployment rate for people aged 16 and over, for the latest three-month period in the ONS Labour market overview published on 18 August 2026, is greater than or equal to 4.6% and less than or equal to 5.2%. If the ONS does not publish a headline 16+ unemployment rate for a three-month period in that release, the forecast is marked withdrawn with that reason rather than resolved either way.
Resolves by
2026-08-18
Model
judgement — reasoning recorded at issue, unchanged since
Ledger id
lxf_external_004
Outcome
TRUE — 4.9%
Evidence, as first published
The UK unemployment rate for people aged 16 years and over was estimated at 4.9% in April to June 2026. (source, published 2026-08-18)
Brier score
0.01 — band midpoint 0.90. 0 is perfect, 0.25 is what always saying 50% scores.
Reasoning recorded at issue — unchanged since, including where it turned out to be wrong: The rate was 4.9% for March to May 2026, down 0.1pp on the previous three-month period and up 0.2pp on a year earlier. Rolling three-month estimates overlap by two months, so consecutive readings move slowly; a 0.3pp band either side of 4.9% covers a very large share of historical one-release movements. The explicit withdrawal clause is not a hedge on the number: the ONS has been transforming its labour market statistics and has changed what it publishes before, and a forecast whose subject stops being published should be recorded as unresolvable rather than scored as a hit or a miss.
If ten forecasts were issued at this band, the band claims roughly 9 of them should come true. Calibration — not certainty — is what the register is scored on.
What would prove this system wrong: a forecast on this register edited after issue, a resolution scored against anything but the source named in advance, or an accuracy claim made before ten outcomes have resolved in a domain. Every one is checkable from the raw ledgerthe full falsification list.