World Cup 2026 Knockout Pairings: What Retained Outputs Show
Publisher: Football Sim Lab · Published · Corrected
Correction, 18 October 2026: The earlier June-labelled estimates could not be attributed to a retained input snapshot. They have been withdrawn. This article now reports counts from named stored output files, not verified forecasts or actual tournament results.
Count pairings before interpreting them
The audit selects a match ID, sorts its two team names and counts that pair once in each output. Order does not matter: France vs Morocco and Morocco vs France are the same pairing at a given match ID.
| Retained file | Match | Most frequent pairing | Count / 1,000 |
|---|---|---|---|
| simulation-results.json | M97 | France vs Morocco | 1,000 |
| simulation-results-v2.json | M97 | France vs Netherlands | 162 |
| simulation-results-v3.json | M97 | France vs Netherlands | 164 |
A fixed-looking pairing is a reason to inspect assumptions
France vs Morocco occupies M97 in every output of the first file. That describes the file, not pre-tournament certainty. Locked scores, bracket inputs or another conditional setup could explain the pattern. The retained metadata is insufficient to distinguish those explanations.
The other files vary at the same slot. Their differences show why a chart should identify its source batch and any fixed outcomes. These tables do not report the actual 2026 bracket or a current matchup forecast.
Trying the question in the current game
Open Group Ranking, fix a score and follow its effect on qualification. Repeat with stored real results disabled. Record which matches were fixed before comparing your outputs. A conditional bracket answers a different question from an entirely simulated tournament.
Evidence and limits
The research audit and downloadable data contain source filenames, SHA-256 checksums, run counts and aggregates. The repository script scripts/generate-research-summary.mjs recomputes the counts. All three batches contain 1,000 outputs.
Original seeds, complete rating inputs, locked results and model commits were not retained alongside these files. Filenames cannot certify a generation date. These outputs cannot establish forecast accuracy or a calibrated real-world probability. Some stored round labels are wrong, so the audit uses match IDs to identify stages. Batches with unknown assumptions must not be pooled.
Read Methodology for the current game engine and How to play for a controlled example. Official tournament information is available from FIFA. These are independent game outputs, not FIFA predictions.