★ Coming July 20, 2026
How well did the model do?
The JetLag XI predictions are a moving target — 10,000 Monte Carlo simulations re-running hourly as real scores come in. After the WC 2026 final, we'll send one email comparing every pre-match prediction to what actually happened.
What's in it
Brier scores per metric
Calibration on pAdvance, pTopOfGroup, pChampion. Were we honest about uncertainty, or did we crow when we shouldn't have?
Hit rate vs confidence band
Of the teams we said had >70% to advance, how many did? >90%? <30%? Where did the model bend reality.
Biggest misses, ranked
The matches where the model was most wrong — and a one-paragraph theory for each. Was it Elo? Travel? Variance?
Day-by-day calibration curves
For any team you care about: how did Spain's championship odds evolve from June 11 to July 19, and did the curve track reality?
Methodology postmortem
What we'd change before doing this again. No hand-waving.
★ One email, no newsletter
Reserve your spot
We won't email between now and the final. No daily digest, no recap, no promotional anything. One email after the final — that's the entire offer.
Get the post-tournament calibration report
After the final on July 19, one email comparing the model's pre-match odds to actual results — Brier scores, hit rates, biggest misses, the honest postmortem. One email, no newsletter.
The data's already being captured
Every day during the tournament, the model saves a full snapshot of every team's predicted odds — advancement, group position, championship, expected points. So far 2 days captured. By the final on July 19, we'll have ~39 days of evolving predictions to compare against actual results. Genuine calibration, not vibes.