Forecast record
Brighton & Hove Albion · Published predictions and official starters
24-hour method: 9.0 of 11 starters matched on average across 1 evaluated team forecasts (1 matches). Early sample; not a reliable estimate of future performance.
24-hour method
Method set on 12 September 2026, before these deadlines.
- Starters matched
- 9 / 11
- Team forecasts
- 11 match
- Awaiting evaluation
- 0
- Excluded
- 5
Early sample; not a reliable estimate of future performance.
Recorded through 1 Nov 2026, 18:57 UTC.
Showing 1–6 of 6 · All records
Scores use the full team and method selection. The record view and page do not change them.
Does the forecast improve on the previous XI?
We compare the forecast with the unchanged eleven from the team's previous Premier League match, recorded together before the same 24-hour deadline.
No paired results yet. Comparison began on 14 October 2026. Earlier results have no retrospectively reconstructed baseline.
Comparison coverage
0 paired · 1 evaluated without a qualified previous XI · 0 awaiting evaluation · 5 excluded.
The comparison uses only paired forecasts. Results above may include forecasts without a recorded baseline. Missing evidence is never scored as zero.
Showing 6 of 6 records in this group.
Evaluated results
Brighton & Hove Albionvs Sunderland · 10 Oct 2026 9 / 11 matched
Forecast recorded .
Last recorded forecast before the 24-hour cutoff. Its timing before the official announcement is not verified.
Started, not predicted2
- Pascal Struijk
- Chema Andrés
Predicted, did not start2
- Lewis Dunk
- Malick Yalcouyé
Starting selections, not substitutions or one-for-one replacements.
Compare all starters
| Player | Forecast | Official XI |
|---|---|---|
| Bart Verbruggen | Starter | Starter |
| Ferdi Kadioglu | Starter | Starter |
| Luka Vuskovic | Starter | Starter |
| Pascal Struijk | Not a starter | Starter |
| Olivier Boscagli | Starter | Starter |
| Pascal Groß | Starter | Starter |
| Chema Andrés | Not a starter | Starter |
| Diego Gómez | Starter | Starter |
| Yasin Ayari | Starter | Starter |
| Maxim De Cuyper | Starter | Starter |
| Charalampos Kostoulas | Starter | Starter |
| Lewis Dunk | Starter | Not a starter |
| Malick Yalcouyé | Starter | Not a starter |
Forecast version & sources
- Forecast recorded
- 24-hour cutoff
- Forecast prepared
- Model checked
- Official source collected
A source collection or observation time is not the official announcement time.
These are the saved starting selections, not a forecast rebuilt with later information.
Evidence at forecast time
Review of missed starters
Predicted, not selected
- Lewis DunkAt forecast: No availability row recordedNo later availability change recorded in this review window.
- Malick YalcouyéAt forecast: No availability row recordedNo later availability change recorded in this review window.
Selected, not predicted
- Pascal StruijkAt forecast: No availability row recordedNo later availability change recorded in this review window.
- Chema AndrésAt forecast: No availability row recordedNo later availability change recorded in this review window.
An absent availability row does not confirm fitness. These records do not establish why the manager chose this XI.
Times and sources for this review
Forecast recorded ; prepared .
Later changes are checked only until kickoff at . The official announcement time is unknown; a later record is not proof that news arrived before the announcement.
Official boundary source: Premier League match centre · Collected . First recorded official XI
9 starters matched: 0 had an availability row at forecast; 9 did not. The same evidence can occur among correct picks.
- Lewis Dunk
- At forecast: No availability row recorded; fitness is not confirmed.
- Malick Yalcouyé
- At forecast: No availability row recorded; fitness is not confirmed.
- Pascal Struijk
- At forecast: No availability row recorded; fitness is not confirmed.
- Chema Andrés
- At forecast: No availability row recorded; fitness is not confirmed.
Workload effects and model errors cannot be inferred from these availability records. They require inputs saved before the match and a separate evaluation.
Review evidence · 1 of 1 evaluated forecast
Across 4 differences in starting status:
- Availability row at forecast
- 0
- Later recorded change
- 0
- Neither recorded
- 4
These evidence groups can overlap. They describe saved records, not the cause of a missed starter.
Not evaluated
Brighton & Hove Albionvs Manchester City · 31 Oct 2026The last answer before the deadline was not an eligible full predicted XI. Not evaluated
The last answer before the deadline was not an eligible full predicted XI.
Brighton & Hove Albionvs Liverpool · 25 Oct 2026The last answer before the deadline was not an eligible full predicted XI. Not evaluated
The last answer before the deadline was not an eligible full predicted XI.
Brighton & Hove Albionvs Crystal Palace · 18 Oct 2026The last answer before the deadline was not an eligible full predicted XI. Not evaluated
The last answer before the deadline was not an eligible full predicted XI.
Brighton & Hove Albionvs Arsenal · 19 Sept 2026The last answer before the deadline was not an eligible full predicted XI. Not evaluated
The last answer before the deadline was not an eligible full predicted XI.
Brighton & Hove Albionvs Coventry City · 13 Sept 2026The last answer before the deadline was not an eligible full predicted XI. Not evaluated
The last answer before the deadline was not an eligible full predicted XI.
Earlier deadlines
These deadlines passed before the method was introduced. Retained public forecasts, not a prospectively specified test.
- Starters matched
- Not evaluated
- Team forecasts
- 00 matches
- Awaiting evaluation
- 0
- Excluded
- 0
No retained team/match observations in this group.
What the record measures
History begins 10 Sept 2026, 20:30 UTC.
Starters matched, not formation, positions, match outcomes or betting returns. Each team counts once per method; corrections replace its score, not its place in the sample. Excluded and pending answers are not counted as zero.
Only actual confirmed public observations are retained. Past forecasts are never rebuilt using today's data. The two methods and the earlier-deadline group are not combined.