Using data from previous tournaments and qualifying, we’ve tasked our AI predictor model with predicting the final score of every game at this summer’s World Cup, starting with the two games taking place on Matchday One.
Our AI predictor combines data from qualifying matches, previous World Cups, bookmakers and more to determine the most likely correct score outcome from the game.
As the tournament progresses, the model will focus more heavily on the games that have already taken place within the tournament and it will look beyond results to input shot, possession and general match momentum data to see where value may lie.
We’ve started for now with a look at the first two games of the tournament.
Mexico vs South Korea
The deeper data actually makes this a more cautious Mexico pick than the odds alone suggest. Mexico were previously priced as a strong favourite by the market, but the advanced numbers from our sample does not show a dominant underlying edge. In fact, South Africa’s recent advanced numbers are stronger on shot generation and control volume, which is why the correct-score pick stays low-scoring rather than moving toward a 2-0 or 3-0 Mexico result.
Those numbers explain why this should not be framed as a Mexico banker. South Africa’s profile is competitive: they took more shots, hit the target more often, created more big chances, completed more passes, and moved the ball into the final third and penalty area more regularly. Their 86.8% pass accuracy and lower loss rate also suggest they are capable of sustaining periods of control rather than simply playing as a low-block outsider.
The reason the predictor still lands on Mexico 1-0 is that the market signal and broader team-strength prior still lean Mexico, while the goal-market shape previously pointed toward a lower-scoring match. But the advanced data materially changes the tone: this is a narrow pick with a draw-risk, not a dominant-favourite prediction. If South Africa reproduce their recent box-entry and shot-volume numbers, they have a realistic path to making the match uncomfortable for the hosts.
AI prediction: Mexico 1-0 South Africa
Get the Squawka Bet team’s Mexico vs South Africa tips and predictions here.
South Korea vs Czech Republic
This is the more tactically interesting matchup because the two teams look strong in different ways. South Korea’s profile is built on control and passing security: they attempted far more passes, completed them at a higher rate, entered the final third regularly, and lost the ball less often relative to their on-ball volume. Czech Republic, however, carried the sharper raw attacking edge, especially for shots, big chances, touches in the box, and penalty-area entries and they carry a specific threat from set-pieces.
The 2-1 South Korea pick comes from that tension. Korea’s case is based on control: 674.3 passes per match, 88.8% pass accuracy, 192.3 final-third passes per match, and only 14.8 losses per 100 on-ball events. That is the profile of a side likely to spend long spells managing the ball and territory. Their unbeaten recent results and strong qualifying standing still matter, but the advanced metrics give a better explanation of how they can win: by controlling rhythm and reducing transition exposure.
Czech Republic’s case for a goal is also strong. They lead Korea in shots per match, shots on target, big chances, box touches, passes into the penalty area, and attacking-third regains. Those are not empty possession numbers; they point to a team that can create high-value moments even if they do not dominate the ball. That is why a Korea clean sheet feels too aggressive. Czechia’s attacking data supports both teams to score, even though Korea remain the marginal pick.
AI prediction: South Korea 2-1 Czech Republic
Get the Squawka Bet team’s South Korea vs Czech Republic tips and predictions here
| Our AI Model’s Prediction | Actual full-time result |
|---|---|
| Mexico 1-0 South Africa | TBC |
| South Korea 2-1 Czech Republic | TBC |
Read more:
