Trust Center
Trust & transparency
How current, reliable and explainable today's data and forecast are.
System status
- Operational
Snapshot freshness
Today's daily Monte-Carlo snapshot is published
latest snapshot 2026-07-20; expected ≥ 2026-07-20
- Operational
Results feed
The live feed knows at least the bundled results
feed finished=102, seed finished=28
- Operational
Daily run
This morning's compute completed without error
today's run idle
- Operational
Daily brief
Recap, stat of the day and movers for the latest snapshot
daily content present for 2026-07-20
These are the same live checks /api/health reports — each row pairs a color, a shape and a label so meaning never rests on color alone.
Freshness
Forecast as of
Monday, July 20, 2026
Tournament day
40 / 39
Latest result reflected
Yes
The current forecast reflects all completed matches.Most recent final: ESP 1–0 ARG, played Sunday, July 19, 2026.
Forecast identity
- Snapshot published
- Monday, July 20, 2026
- Simulations
- 250,000
- Methodology version
- v1.4
- Input data effective
- Monday, July 20, 2026
- Includes latest result
- Yes
Corrections ledger
| Date | Affected | Previous | Corrected | Reason |
|---|---|---|---|---|
No public corrections recorded this tournament. When a published figure is amended, it is logged here with what changed and why — the legend above shows the structure each entry would take. | ||||
Methodology & changelog
Read the methodology →The full engine write-up — ratings, the goal model, determinism and known limitations — lives on the methodology page. Recent product changes:
- Jun 2026Editorial redesign — The AlmanacA luxe, magazine-grade rebuild of every public surface.
- May 2026Match Leverage addedPer-fixture swing scores quantify how much each match moves the field.
- Apr 2026My Cup personalizationFollow nations to assemble a personalized tournament view.
- Mar 2026Trust Center launchedThis page — system status, freshness and forecast provenance in one place.
Calibration
The forecast is a deterministic Monte-Carlo engine: every morning it fixes the results that have actually happened and simulates the remainder of the tournament 250,000 times, counting how often each outcome occurs. Because probabilities are read straight off those counts, they are calibrated by construction — a 30% chance happens in close to 30% of runs.
The honest caveats: sample sizes for rare events (a specific final pairing, an unlikely run) are small, so those figures carry real noise; and the model cannot see injuries, suspensions, tactics or form. These are forecasts, not guarantees — a strong favorite still loses often enough to matter.