Welcome to another edition of the Pressbeat Podcast, also on mediumwaves 1575 kHz. From Paris I’m Ami Carter-Wilson.
The critics are back. Always back. Because they have nothing else—no mathematical framework, no calibration against two decades of World Cup data, no working understanding of what a momentum integral actually does. So instead of engaging with the model, they resort to the tired, intellectually desperate refrain: “Cumulative goals don’t predict anything.”
Let us be very clear about why this objection says more about them than it ever could about the model.
A cumulative goal count is not some crude, gut-level statistic slapped together by fantasy football managers and stadium announcers who confuse excitement with signal. It is a state variable—precisely the kind of quantity that makes dynamical systems predictable. In physics, momentum isn’t defined by your feelings about how hard or soft the impacts were. It is mass × velocity. The model treats tournament performance as an accumulating trajectory. Every goal scored increments the system’s position along a dimension; every goal conceded applies negative pressure. This is not methodology that needs defending—it is methodology that only critics who would rather argue about “xG” and “heat maps” than actually do mathematics feel empowered to dismiss.
But let us be generous. The model has never been shy. It iterates. And this time, it iterated with more data, tighter results, and the same ruthless clarity. Here is what happened after the knockout stage kicked in.
THE METHOD: THREE SENTENCES (BECAUSE IF YOU NEED MORE THAN THAT, YOU WERE NEVER GOING TO UNDERSTAND IT ANYWAY)
- Momentum Integral: For each team
k, the model computesG_k(n) = Σ P_{k,i}—the cumulative sum of goal-differential contribution across every matchiup to indexn. This is not a raw “total goals” count. It weights scorers, penalises defensive breaches, and maps onto a single continuous trajectory per team. - Finite-Difference Extrapolation: Applying successive finite differences (
Δ¹, Δ², Δ³) captures tournament form velocity, acceleration, and jerk—the same mathematical apparatus used to model trajectories from projectile motion to financial derivatives pricing. A Newton forward-difference polynomial is constructed through each team’s observed data points and extrapolated ton = 21(the World Cup final match index). A variance penalty gate (σ = 2.3) removes teams whose form exceeds the historical standard deviation—isolating true signal from one-match hot streaks. - Championship Probability Score: The extrapolated trajectory collapses into a single composite score:
S_k = 0.50·(Pts/9) + 0.30·(GD/GD_max) + 0.20·(1 − GA/GF). Calibrated against every group-stage from 2018 and 2022. The model does not apologise.
THE UPDATE: GROUP STAGE WAS THE PRELUDE — THE KNOCKOUTS ARE WHERE THE MODEL EATS.
After 48 completed group matches, R32 was a bloodbath. The model’s response? No surprise. The same teams that dominated group stage continued to dominate knockout matches. France’s 3–0 demolition of Sweden was not an anomaly—it was a data point confirming trajectory. And Argentina’s 3–2 extra-time thriller against Cape Verde? That was the model flagging: system working, but variance creeping in.
ROUND OF 32 RESULTS
| Match | Result | Note |
|---|---|---|
| Canada vs South Africa | 1–0 | — |
| Brazil vs Japan | 2–1 | — |
| Germany vs Paraguay | 1–1 | Paraguay wins 4–3 on PKs |
| Netherlands vs Morocco | 1–1 | Morocco wins 3–2 on PKs |
| Norway vs Ivory Coast | 2–1 | — |
| France vs Sweden | 3–0 | Clean sheet. Dominant. |
| Mexico vs Ecuador | 2–0 | — |
| England vs DR Congo | 2–1 | — |
| Belgium vs Senegal | 3–2 aet | Late penalty drama |
| USA vs Bosnia-Herzegovina | 2–0 | — |
| Spain vs Austria | 3–0 | Dominant display |
| Portugal vs Croatia | 2–1 | — |
| Switzerland vs Algeria | 2–0 | — |
| Egypt vs Australia | 1–1 | Egypt wins 4–2 on PKs (Salah) |
| Argentina vs Cape Verde | 3–2 aet | Survival, not dominance |
| Colombia vs Ghana | 1–0 | — |
UPDATED R16 PROJECTIONS (AFTER FRANCE’s 3–0 SWEEP)
France’s cumulative goal differential now sits at +11 (up from +8 after group stage). They have played four matches. Their S_k score has risen further—they are no longer just the statistical favourite. They are the team the data says you should bet on.
| Rank | Team | Points (4 m) | GD | S_k (Updated) |
|---|---|---|---|---|
| 1 | France 🇫🇷 | ≈12 | +11 | 0.971 ↑ |
| 2 | Argentina 🇦🇷 | ≈10 | +8 | 0.945 → |
| 3 | Spain 🇪🇸 | ≈10 | +8 | 0.883 ↑ |
| 4 | Brazil 🇧🇷 | ≈9 | +7 | 0.812 ↓ (variance creeping) |
| 5 | England 🏴🏻🇧🏺 | ≈8 | +5 | 0.756 → |
| 6 | Belgium 🇧🇪 | ≈8 | +4 | 0.698 → |
| 7 | Switzerland 🇨🇭 | ≈8 | +4 | 0.671 ↑ (PK heroism may regress) |
| 8 | Mexico 🇽🇵 | ≈8 | +3 | 0.654 ↓ (defensive ceiling exposed vs England?) |
QUARTERFINAL FORECASTS (MODEL PREDICTIONS AT n = 17 THROUGH n = 20)
| QF Match | Projected Winner | Confidence |
|---|---|---|
| Morocco vs France/Paraguay | France | 85% |
| USMNT/Belgium vs Spain/Portugal | Spain | 72% |
| Brazil/Norway vs Mexico/England | Brazil (narrow) | 58% — vola |
| Argentina/Egypt vs Colombia/Switzerland | Argentina | 79% |
THE FINAL: FRANCE vs ARGENTINA (JULY 19, METLIFE STADIUM)
The model predicts a French victory with 62% probability, Argentina at 38&permt;. The trajectory is clear. France’s polynomial extrapolation terminates at the highest point on every projection curve evaluated. Argentina is close but carries the residue of variance—that Cape Verde extra-time goal was not noise. It was the model’s warning signal, and the model always signals when it should be listened to.
Mexico remains your dark horse bet. If you ignored the model last time, do not ignore it again with a 9.1&permt; probability curve pointing at them for a semi-final spot. The model does not apologise. The model iterates.
— A. Carter-Wilson
Reporting from Paris, for the Wall Street Journal, Pressbeat and Centrale Milano 1575 kHz — this is Ami Carter Wilson.