Free tournament management · FIDE 2026 rules. Open app →
Technical evidence
How OpenPairing is verified
Promoted results, complete denominators and explicit boundaries. Quarantines, incomplete attempts and negative controls remain separate from accepted results.
| Campaign | Coverage | Tournaments | Verified rounds | Result |
|---|---|---|---|---|
| p50 | 384/384 shard | 30,000 | 768,324 | PASS |
| p80 | 256/256 shard | 20,000 | 660,000 | PASS |
| p150 | 128/128 shard | 2,000 | 182,000 | PASS |
| Acceleration epoch 5 | 64/64 shard | 1,000 | 11,000 | PASS |
| p300 synthetic campaign | 64/64 shard | 200 | 12,200 | PASS |
| p600 pilot | 1/1 case | 1 | 31 | PASS |
The 53,201 synthetic inputs and 1,633,555 exact rounds are the publishable Gacrux synthetic-campaign selection. The separate historical replay covers 122,008 decodable tournaments and 2,220,576/2,220,579 reference-exact rounds; all three discrepancies are classified. OpenPairing does not redistribute the TRFs.
The historical corpus is not presented as a perfect-only result: 122,006 tournaments are reference-exact; across ten misc cases, 378/381 rounds match the historical reference and three divergent pairings instead match Gacrux-computed pairs. The deviations remain in the denominator and are preserved and explained in the reports.
Read the historical verification →Chain of trust
- Source and rules are frozen.
- Runtime identity and hashes are verified.
- Report and classification are promoted atomically.
- Rebuildable aggregates and preserved negative controls.
Declared limitations
PASS means that a specific campaign met its acceptance criteria. The 200-tournament p300 synthetic campaign does not close the separate 547-tournament capacity gate above 256 players. PASS does not mean FIDE approval and does not replace complete-product tests, shadow validation or TEC decisions.
Report reproducible cases, propose synthetic fixtures or review the tools and methodology on GitHub. Do not include personal data or non-redistributable tournaments.
Open an issue →