Forecast record
Arsenal · Published predictions and official starters
Earlier deadlines: 11.0 of 11 starters matched on average across 1 evaluated team forecasts (1 matches). Early sample; not a reliable estimate of future performance.
No retained record for this team, match and method. Other results below are not a substitute.
24-hour method
Method set on 12 September 2026, before these deadlines.
- Starters matched
- Not evaluated
- Team forecasts
- 00 matches
- Awaiting evaluation
- 1
- Excluded
- 4
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 · 0 evaluated without a qualified previous XI · 1 awaiting evaluation · 4 excluded.
The comparison uses only paired forecasts. Results above may include forecasts without a recorded baseline. Missing evidence is never scored as zero.
Showing 5 of 5 records in this group.
Awaiting evaluation
Arsenalvs Everton · 24 Oct 202624-hour cutoff: · The deadline has passed; official lineup evidence is still missing. Awaiting evidence
The deadline has passed; official lineup evidence is still missing.
Not evaluated
Arsenalvs Liverpool · 1 Nov 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.
Arsenalvs Nottingham Forest · 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.
Arsenalvs Leeds United · 10 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.
Arsenalvs Brighton & Hove Albion · 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.
Earlier deadlines
These deadlines passed before the method was introduced. Retained public forecasts, not a prospectively specified test.
- Starters matched
- 11 / 11
- Team forecasts
- 11 match
- Awaiting evaluation
- 0
- Excluded
- 0
Early sample; not a reliable estimate of future performance.
Showing 1 of 1 records in this group.
Evaluated results
Arsenalvs Sunderland · 12 Sept 2026 11 / 11 matched
Forecast recorded .
Last recorded forecast before the 24-hour cutoff. Its timing before the official announcement is not verified.
Recorded before the 24-hour method was introduced.
All 11 predicted starters were in the official XI.
Compare all starters
| Player | Forecast | Official XI |
|---|---|---|
| David Raya | Starter | Starter |
| Ben White | Starter | Starter |
| Ezri Konsa | Starter | Starter |
| Gabriel Magalhães | Starter | Starter |
| Riccardo Calafiori | Starter | Starter |
| Declan Rice | Starter | Starter |
| Myles Lewis-Skelly | Starter | Starter |
| Bukayo Saka | Starter | Starter |
| Martin Ødegaard | Starter | Starter |
| Christos Tzolis | Starter | Starter |
| Kai Havertz | Starter | 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.
The method was set after this match's cutoff. This retained forecast is not a prospectively specified test.
These are the saved starting selections, not a forecast rebuilt with later information.
Evidence at forecast time
All 11 starters matched. There are no missed starters to review.
Review evidence · 1 of 1 evaluated forecast
Across 0 differences in starting status:
- Availability row at forecast
- 0
- Later recorded change
- 0
- Neither recorded
- 0
These evidence groups can overlap. They describe saved records, not the cause of a missed starter.
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.