Checking our work

Do the probabilities come true?

When the model says a team has a 70% chance, that team should win about seven times in ten. This page checks whether it does. Every 2026 World Cup match is graded the moment it's played, and the numbers below ask one question: are the probabilities honest, not just confident? (The technical name for this is calibration.)

Track record

Proven on past tournaments

The short version: when the model says 70%, it happens about 70%. Across these 987 matches its stated chances landed within ~5.6 points of reality, and on average it rated the actual result about 35% more likely than a blind 1-in-3 guess would.

The 2026 tracker below stays empty until the first match kicks off. So to show the model has been tested, not just described, we ran it against tournaments whose results you already know. For each one, the model was rebuilt exactly as it stood the day before kickoff, then graded on every match — it never sees the result it is being marked on. That is what “graded out of sample” means: no peeking, no hindsight.

Each tournament is scored by the production model reconstructed as it stood the day before the tournament's first match: Dixon-Coles and Hierarchical Poisson refit on matches strictly beforehand, Elo rolled forward to each match, and the tournament-tier calibration layer refit on the 24 months of matches before the cutoff. No data from the tournament, or any later match, touches any layer of the fit.

Graded across 987 matches at 24 major tournaments (2014–2024)

0.572

How close the forecasts landed to reality. Lower is better; blind 1-in-3 guessing scores ≈ 0.667.

1.000

Like Brier, but overconfidence is punished harder. Lower is better; blind guessing scores ≈ 1.099.

5.6pp

Does “70%” really mean 70%? The average gap between the two. Lower is better.

Tournament by tournament

One row per tournament: the model rebuilt as it stood the day before it began, then graded on every match through the final (Brier — lower is better, blind 1-in-3 guessing is 0.667). A few thin, early editions sit above that line; the honest measure is all of them pooled, in the box above.

TournamentHostMatchesBrier
Copa América 2024United States320.522
Euro 2024Germany510.613
AFCON 2024Ivory Coast520.651
Asian Cup 2024Qatar510.515
Gold Cup 2023United States310.566
World Cup 2022Qatar640.611
AFCON 2022Cameroon520.686
Gold Cup 2021United States310.341
Copa América 2021Brazil280.481
Euro 2021England510.554
AFCON 2019Egypt520.546
Gold Cup 2019United States310.405
Copa América 2019Brazil260.542
Asian Cup 2019United Arab Emirates510.496
World Cup 2018Russia640.569
Gold Cup 2017United States250.456
AFCON 2017Gabon320.642
Euro 2016France510.668
Copa América 2016United States320.502
Gold Cup 2015United States260.755
Copa América 2015Chile260.686
AFCON 2015Equatorial Guinea320.795
Asian Cup 2015Australia320.434
World Cup 2014Brazil640.565

Reliability diagram

Read it like this: each dot is a group of similar forecasts. left-to-right is what the model predicted, bottom-to-top is how often it actually happened. When the model says 70% and that happens about 70% of the time, the dot sits on the dashed line: perfect calibration. The closer the dots hug the line, the more honest the probabilities; bigger dots mean more matches in that group.

Reliability diagramReliability diagram: predicted probability on the x-axis, observed frequency on the y-axis, binned in deciles across [0, 1]. Closer to the identity line means better-calibrated.0.000.250.500.751.000.000.250.500.751.00n=327n=478n=911n=329n=269n=217n=184n=134n=73n=39predicted probabilityobserved frequency

Brier by favourite confidence

Matches grouped by how confident the model's favourite was (its biggest of the home / draw / away probabilities) — so you can see whether it is as reliable on toss-ups as on heavy favourites.

Favourite confidenceMatchesBrier
P_fav < 40%810.649
P_fav 40-60%4760.633
P_fav 60-80%3180.512
P_fav >= 80%1120.428
  • Out-of-sample: the calibration layer is refit per tournament on pre-tournament data, so these numbers do not reuse the live shipped calibrator (which has seen these results).
  • The uniform 1/3 forecast scores a Brier of 0.667; lower is better. Major-tournament football is high-variance, so a strong model still sits well above a league-season Brier.
  • Calibrated and uncalibrated metrics are reported on the same fixtures so the calibration layer's effect is visible.

Built 2026-05-30 · model 1.0.0 · calibration layer refit on the 24 months before each tournament.

2026 World Cup, live tracker

Across 95 graded forecasts so far, the model's overall Brier score is 0.497 (lower is better; blind 1-in-3 guessing scores ≈ 0.667). The reliability diagram below plots what the model predicted against how often it actually came true. The closer to the diagonal, the more honest the probabilities.

Overall, across 95 scored matches

0.497

How close the forecasts landed to reality. Lower is better; blind 1-in-3 guessing scores ≈ 0.667.

0.837

Like Brier, but overconfidence is punished harder. Lower is better; blind guessing scores ≈ 1.099.

0.165

Does “70%” really mean 70%? The average gap between the two. Lower is better.

Rolling Brier, 10-match window

The model's accuracy over its most recent 10 matches, re-figured after each game. Lower is better; the dashed line is what blind 1-in-3 guessing would score, so anything below it is real skill.

Rolling Brier line chart, 10-match window across 95 scored matchesRight-aligned rolling mean of the per-match 3-class Brier score over the last 95 scored World Cup matches. Lower is better; the dashed line at 0.667 is the uniform-1/3 baseline.0.000.250.500.751.00match index (1 → 95)

Per-matchday breakdown

DateMatchesBrierLog loss
2026-06-1120.4040.723
2026-06-1220.8171.280
2026-06-1340.7711.178
2026-06-1440.5951.008
2026-06-1541.1481.632
2026-06-1640.1970.438
2026-06-1740.6401.005
2026-06-1840.4260.739
2026-06-1940.4630.787
2026-06-2040.5310.854
2026-06-2140.6030.928
2026-06-2240.2660.529
2026-06-2340.4330.706
2026-06-2460.4100.750
2026-06-2560.5060.868
2026-06-2660.3880.704
2026-06-2760.4340.763
2026-06-2920.6120.996
2026-06-3030.5000.866
2026-07-0130.4780.830
2026-07-0230.3270.622
2026-07-0330.4600.796
2026-07-0410.2250.470
2026-07-0620.4430.784
2026-07-0720.5410.909
2026-07-0910.5730.960
2026-07-1010.2800.564
2026-07-1120.2500.517

Updated 2026-07-14.

Reliability diagram

Read it like this: each dot is a group of similar forecasts. left-to-right is what the model predicted, bottom-to-top is how often it actually happened. When the model says 70% and that happens about 70% of the time, the dot sits on the dashed line: perfect calibration. The closer the dots hug the line, the more honest the probabilities; bigger dots mean more matches in that group.

Reliability diagramReliability diagram: predicted probability on the x-axis, observed frequency on the y-axis, binned in deciles across [0, 1]. Closer to the identity line means better-calibrated.0.000.250.500.751.000.000.250.500.751.00n=27n=46n=95n=33n=21n=25n=18n=14n=4n=2predicted probabilityobserved frequency

Brier by competition

SegmentMatchesBrierΔ vs overall
World Cup 2026950.497+0.000

Brier by tournament stage

SegmentMatchesBrierΔ vs overall
Group stage720.516+0.019
Round of 32150.449-0.048
Round of 1640.492-0.005
Quarter-final40.338-0.159
Semi-final
Third-place play-off
Final

Brier by favourite confidence

SegmentMatchesBrierΔ vs overall
P_fav < 40%110.669+0.172
P_fav 40-60%460.540+0.043
P_fav 60-80%320.373-0.124
P_fav >= 80%60.513+0.016

What the three numbers mean

Think of a weather forecaster. Anyone can say “70% chance of rain.” The good ones are actually right about 70% of the time when they say it. These three numbers check the model the same way.

  • Brier score: were the probabilities close to reality? For every match we measure how far the forecast landed from what actually happened, then average it. A perfect crystal ball scores 0; blindly guessing 1-in-3 every time scores about 0.667. Lower is better.
  • Log loss: the same idea, but overconfidence is punished hard. Call something nearly certain and then get it wrong, and this number jumps. It is the metric that keeps the model humble. Blind guessing scores about 1.099. Lower is better.
  • ECE: does “70%” really mean 70%? We gather up every “about 70%” forecast and check how often those things actually happened. The average gap, across every confidence level, is the ECE. A few percentage points means the stated chances can be taken at face value. Lower is better.

The first two reward being right and bold; the last is the honesty check. A model can look impressive and still overstate its confidence. Measuring all three together is what catches that.

Want the machinery underneath, the component models and the held-out test behind these numbers? It is all on the methodology page.

The metrics on this page are the model's own calibration, scored against observed match outcomes: the predicted probabilities compared to what actually happened on the pitch.

In-tournament calibration · onthepitch · onthepitch