Football contains something algorithms cannot fully capture: unpredictability.
As the 2026 World Cup unfolds, a financial firm has done what humans have always done before great contests: attempted to impose order on uncertainty. Credicorp Capital's mathematical model, built from Elo ratings, Monte Carlo simulations, and the wisdom of prediction markets, places Spain as the tournament's most likely champion while offering Colombia a modest but meaningful three percent chance at glory. The exercise reminds us that probability is not destiny — it is merely the shape our hopes and fears take when we dress them in numbers.
- Spain emerges as the model's dominant force, its 33.7% title probability nearly doubling that of second-place France, signaling a statistical gulf between the favorites and the rest of the field.
- Colombia's inclusion among the top ten contenders creates genuine excitement, even as the algorithm projects their run ending in the quarterfinals against the very team favored to win it all.
- The model's predicted bracket — Spain over France in the semis, England over Argentina, then Spain defeating England in the final — constructs a narrative that feels both plausible and fragile.
- Rafael Castellanos, the analyst behind the work, openly acknowledges what 50,000 simulations cannot resolve: football's irreducible human unpredictability, which Colombia's talent is more than capable of exploiting.
When the 2026 World Cup began, Credicorp Capital released a predictive model that combined Elo ratings, FIFA rankings, Poisson score distributions, and over 50,000 Monte Carlo simulations — then cross-referenced the results against international betting markets like Polymarket and Kalshi. The conclusion was striking in its clarity: Spain, at 33.7%, stood as the overwhelming favorite, nearly twice as likely to win as France at 18.5%, and more than three times as likely as England at 11.5%. Argentina followed at 9.3%, with Portugal, Brazil, and the Netherlands completing the traditional hierarchy of power.
For Colombian fans, the model offered a bittersweet projection. The squad was expected to finish second in Group K behind Portugal, then defeat Croatia in the Round of 16 — a result that would mark genuine progress. The quarterfinals, however, would bring Spain, and there the algorithm saw Colombia's tournament ending. It was a respectable arc, not a dismissal.
The model's full bracket envisioned a tournament that narrowed, as great tournaments do, toward a final between two of football's most storied nations. France would fall to Spain in the semifinals; Argentina, the defending champions, would fall to England. The final would belong to Spain.
Yet Rafael Castellanos, who leads asset management at Credicorp Capital, was careful to frame the work honestly. The model is a probabilistic tool, not a prophecy. Football, he noted, carries within it something numbers cannot fully hold — and teams like Colombia carry enough talent to remind the world of that fact.
As the 2026 World Cup began, a financial analysis firm called Credicorp Capital released a mathematical model designed to predict the tournament's outcome. The results placed Spain as the overwhelming favorite to win it all, with a 33.7 percent chance of lifting the trophy. Colombia, meanwhile, landed in the conversation as one of ten teams with a realistic path to the championship, though with more modest odds: three percent.
The model combined multiple statistical approaches to reach these conclusions. Credicorp Capital layered together the Elo rating system, FIFA's official rankings, Poisson distributions to forecast match scores, and then ran more than 50,000 complete tournament simulations using Monte Carlo methods. To ground the results in real-world sentiment, the researchers cross-checked their algorithmic output against international prediction markets like Polymarket and Kalshi, where bettors were already pricing their own expectations into the odds.
The full ranking told a familiar story with one notable twist. France came second at 18.5 percent, England third at 11.5 percent, and Argentina fourth at 9.3 percent. Portugal, Brazil, and the Netherlands rounded out the traditional powerhouses. But Spain's 33.7 percent probability stood apart—nearly double France's chances and more than triple England's. The model saw something in the Spanish team that suggested a clear path through the field.
When the simulation played out the group stage, Colombia advanced from Group K as the second-place team behind Portugal. The model then projected the Colombian squad would defeat Croatia in the Round of 16, a respectable result that would send them into the quarterfinals. There, however, the run ended. Spain awaited, and according to the algorithm, the Spanish would eliminate Colombia and continue their march toward the final.
The model's predicted knockout rounds painted a specific vision of the tournament's arc. France would beat Germany in the Round of 16, the Netherlands would dispatch South Korea, Brazil would overcome Norway. England would eliminate Mexico, Belgium would beat the United States, and Argentina would get past Turkey. By the semifinals, the model saw France facing Spain and England facing Argentina. Spain would advance past France, while England would overcome the defending world champions. The final would pit Spain against England, with the Spanish prevailing. France would then beat Argentina in the third-place match.
Rafael Castellanos, who leads asset management at Credicorp Capital, offered a necessary caveat. The model, he explained, provides a statistically robust approximation of what might happen. But football contains something that algorithms cannot fully capture: unpredictability. Teams like Colombia, he noted, possess enough talent and potential to disrupt any forecast built from numbers and equations alone. The model is a tool for understanding probability, not a crystal ball. What happens on the pitch remains, in the end, a human affair.
Citas Notables
Selecciones como Colombia cuentan con el potencial suficiente para alterar cualquier pronóstico construido a partir de algoritmos y modelos matemáticos.— Rafael Castellanos, Managing Director Asset Management, Credicorp Capital
La Conversación del Hearth Otra perspectiva de la historia
Why did Credicorp Capital build this model in the first place? What's the business case for predicting a World Cup?
Financial firms use predictive models to understand risk and opportunity across markets. A World Cup model lets them test their statistical methods and also speaks to clients who care about these outcomes—whether for betting, sponsorship decisions, or just intellectual curiosity about how data can illuminate sports.
Spain at 33.7 percent seems extraordinarily high. What does the model see in them that makes them such a clear favorite?
The model weighs historical performance, current rankings, and player strength. Spain has been consistently excellent, and the Elo system rewards sustained success. But the model is only as good as its inputs—if Spain's squad has weaknesses the numbers don't capture, the forecast could be wrong.
Colombia gets three percent. That's real, though. Not a rounding error.
Right. Three percent means the model sees a genuine path: win the group or finish second, beat a weaker opponent in the Round of 16, then face a tougher team in the quarterfinals. It's not likely, but it's plausible. The model isn't saying Colombia can't win; it's saying the odds are long.
What does Castellanos mean when he says football has an unpredictable component?
He's acknowledging the limits of the model itself. A player gets injured. A goalkeeper has the game of his life. A team finds unexpected chemistry. These things don't live in the algorithm. Colombia could absolutely disrupt this forecast if the conditions align right.
So why publish the model at all if you know it could be wrong?
Because it's useful to think through the probabilities, even if they're not perfect. It gives you a baseline. And it's honest about what it is—a statistical exercise, not destiny. The model is a conversation starter, not a prediction written in stone.