Public forecast 005 · permanent record

Will registered company insolvencies in England and Wales for July 2026 be between 1,550 and 2,150 inclusive?

WITHDRAWN — kept on the record with its reason; withdrawals are never deleted.

Question
Will registered company insolvencies in England and Wales for July 2026 be between 1,550 and 2,150 inclusive?
Probability band
80–90%
Issued
2026-08-14 00:00 UTC
Data cut-off
2026-08-14 00:00 UTC
Frozen criterion
Resolves TRUE if the number of registered company insolvencies in England and Wales in July 2026, as first published by the Insolvency Service in the monthly company insolvency statistics on 18 August 2026, is greater than or equal to 1,550 and less than or equal to 2,150. Scored against the figure as first published.
Original scheduled resolution date (historical — the entry was withdrawn and will not be scored)
2026-08-18 08:00 UTC
Origin
Editorial forecast — operator judgement, scored on the editorial board; not evidence for the decision models — reasoning recorded at issue, unchanged since
Ledger id
lxf_external_005
Sealed receipt
6bad8cdc4d7010c0f7ede59a93af4dd6325b7d17273ee5fa7d5663f28cb25103 — SHA-256 over the entry’s frozen fields at issue; how to check it
Correction, recorded after issue (band and receipt unchanged)
Added 2026-08-14, before publication. The reasoning cites 1,849 for May 2026, which is the figure quoted in the June commentary. The May commentary as first published gives 1,868. Both are official — monthly insolvency counts are revised — which is exactly why this forecast is scored against the figure AS FIRST PUBLISHED. The band is unchanged: 1,845 remains the correct June base it was set from.
Critique on the record
An external review computed a 6.88% standard deviation of one-month changes over Dec 2025 to Jun 2026 and concluded this band spans about -2.6 to +2.2 sigma, i.e. nearer 98% coverage than the 80-90% stated. That critique is probably right and this forecast is probably under-confident. It is being left as issued rather than tightened: n=6 is a thin basis for a sigma, and a band widened by caution and then narrowed after someone points it out is a band chosen to look good. It will be scored as written, and if it resolves comfortably inside, that is evidence of under-confidence which belongs in the calibration record.
Withdrawn
2026-08-14 18:58 UTC
Why it was withdrawn
Issued and PUBLISHED on lexun.co.uk on 14 August 2026, then withdrawn the same day. The due date was set to 2026-08-18T08:00:00Z, which is 09:00 BST. The Insolvency Service publishes its monthly statistics at 09:30 BST — verified from the GOV.UK content API, which gives first_published_at 2026-07-17T09:30:04+01:00 for the June release. The forecast was therefore unresolvable at its own due date, thirty minutes before the only source that can settle it. Due dates are immutable, so this is corrected by supersession rather than edit. Superseded by lxf_external_009, which carries the same question, band and criterion with a reachable due date.
Superseded by
forecast 009 — the replacement names this entry; this one stays on the record, void and unscored
Reasoning recorded at issue — unchanged since, including where it turned out to be wrong: There were 1,845 registered company insolvencies in June 2026 and 1,849 in May 2026, with June 10% below the 2,048 recorded in June 2025. The band spans roughly 16% either side of the two most recent months, which are themselves almost identical. Monthly counts are affected by the number of working days and by administrative processing, so single-month swings of ten per cent are ordinary; swings beyond sixteen per cent are not. This is the entry in the cohort closest to LEXUN's own subject matter, since the runway model exists to forecast whether a business is one of these numbers.
If ten forecasts were issued at this band, the band claims roughly 8 or 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 ledger — the full falsification list.