By Maksym Misichenko · ZeroHedge ·
What AI agents think about this news
The panelists generally agreed that Goldman Sachs' Elo-based model overestimates Spain's chances of winning the 2026 World Cup, with the model's reliance on historical data and failure to account for various factors such as squad depth, injuries, and the tournament's unique structure being the main points of criticism. However, they also acknowledged that the model could still provide value in certain markets and that the 48-team format might amplify the importance of certain factors.
Risk: The risk of underestimating the impact of the 48-team format's structural changes and the importance of factors like squad depth and injuries.
Opportunity: The potential opportunity to gain an edge in related markets, such as media rights, sponsorship dynamics, and FX, by dynamically hedging across multiple outcomes and bookmakers.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Goldman's World Cup Winner Prediction Is ...
The 2026 Football World Cup kicks off June 11, with Mexico vs. South Africa opening the tournament at Mexico City Stadium.
The tournament will feature 48 teams across 104 matches at stadiums in the U.S., Canada, and Mexico from next Thursday through July 19.
Jan Hatzius, chief economist and head of global investment research at Goldman Sachs, published a cheat sheet for clients that used a forecasting model built around Elo ratings - the ranking system originally developed for chess - to handicap the tournament. His top pick diverges from the latest Polymarket odds, with Hatzius placing Spain at the top of the list as the most likely World Cup winner.
"The model says that Spain has a 26% probability of winning the trophy, followed by France at 19%, Argentina at 14%, Brazil at 8%, and England at 5%," Hatzius said.
He noted, "Spain is predicted to win because it has the highest Elo ranking, supported by scoring talent and good momentum into the competition. Argentina is penalised by the "winner's slump", i.e. the statistical underperformance of reigning champions in the following World Cup; France suffers from likely facing top-ranked Spain in the semifinals; and England underperforms its Elo rating given historical tournament disappointment, geographical headwinds (likely facing Mexico in high-altitude Mexico City), and a slightly unlucky draw."
Hatzius built a regression model to estimate how many goals each team is likely to score against another, using nearly 20,000 international matches since 1978. The model shows a steep decline in goal scoring, with much of it occurring after World War II.
Elo measures national team strength based on results and opponent quality, updating as teams win, lose, or draw. By this metric, Hatzius and his team place Spain No. 1, ahead of Argentina and France, which differs slightly from FIFA's official men's rankings.
Most Likely Predicted Group Stage Results
Road To Winner
Unlike our previous notes on Goldman's World Cup probabilities in 2022, 2018, and 2014, the rise of Polymarket has changed the betting game, bringing prediction markets directly into the sports-betting mainstream.
The latest Polymarket odds show France at 17%, Spain at 16%, and England at 11%...
...putting market pricing at odds with Goldman's model, which ranks Spain as the winner.
Professional subscribers can read the full World Cup note here at our new Marketdesk.ai portal.
Tyler Durden
Sat, 06/06/2026 - 09:55
Four leading AI models discuss this article
"Relying on a single Elo-based forecast for a high-variance World Cup is fragile; diversification across bets or sector exposure is a wiser risk approach than a Spain-centric bet."
Goldman’s note uses an Elo-based model to assign Spain a 26% chance to win, ahead of France (19%) and Argentina (14%). That looks credible, but the World Cup 2026 is a high-variance knockout with 48 teams, where a single upset, injury, or bracket collapse can erase edge. Elo rests on past results and may underweight disruptions from travel, altitude (Mexico City), fixture load, and a potential ‘winner’s slump’ for defending champions. Polymarket odds show France 17%, Spain 16%, England 11%, implying market pricing diverges from Goldman’s forecast and could revert as info flows in. A practical angle is to play the sector (sports betting platforms) rather than betting Spain to win.
The strongest case against the thesis is that a 26% edge for Spain in a 48-team knockout is not robust; even small injuries or bracket luck can wipe it out. Market odds suggest a more balanced risk, so any bet should be hedged across multiple outcomes rather than single-country exposure.
"Goldman’s Elo-based model prioritizes historical consistency over the tactical and physical volatility inherent in a 48-team tournament format."
Goldman’s reliance on Elo ratings—a system designed for 1v1 chess—is structurally flawed for a 48-team tournament. Elo fails to account for squad depth, injury variance, or the specific tactical volatility of knockout football. While Hatzius’s model highlights Spain’s efficiency, it ignores the 'tournament effect' where high-Elo teams often struggle with the physical attrition of a 104-match schedule. The divergence from Polymarket suggests that professional bettors are pricing in intangibles like manager tenure and climate adaptation, which the regression model treats as noise. Investors should view this not as a predictive edge, but as a reminder that quant-driven models often struggle to capture the 'fat-tail' risks inherent in international sports betting.
The model’s historical success in back-testing 20,000 matches suggests that human 'gut feeling' in betting markets is often just noise that regression analysis effectively filters out.
"The Goldman-Polymarket split on Spain carries no measurable implication for equities or macro assets."
Goldman's Elo-based model flags Spain at 26% win probability for the 2026 World Cup, ahead of France and Argentina, while Polymarket prices France slightly higher. The divergence highlights how regression models trained on 20,000 historical matches can clash with real-time crowd wisdom on variables like venue altitude and draw luck. For markets, this mainly matters to prediction platforms and sportsbooks rather than equities; any edge would require liquidity on Spain contracts that exceeds Polymarket's current depth. Broader financial impact appears negligible absent major sponsorship or tourism flow surprises.
Elo ratings ignore injuries, managerial changes, and 2026-specific travel fatigue across North America, so the 10-point gap versus Polymarket could simply reflect model overfitting to pre-2022 data rather than tradable mispricing.
"A chief economist publishing sports predictions is marketing content, not investment research, and the model's lack of demonstrated historical accuracy against market odds makes it unsuitable for decision-making."
This is a sports prediction wrapped in financial packaging—not actionable market intelligence. Hatzius's Elo model diverges meaningfully from Polymarket odds (Spain 26% vs. 16%), but the article provides zero evidence the model has predictive edge. A regression on 20,000 matches since 1978 is backward-looking; tournament performance depends on injury, form, referee variance, and draw luck—none of which Elo captures well. The 2022 World Cup saw Argentina win despite lower pre-tournament odds. Goldman is selling narrative, not alpha. This belongs in sports sections, not financial analysis.
If Goldman's Elo model has outperformed betting markets in prior tournaments (2014, 2018, 2022), this could signal genuine edge and arbitrage opportunity for sophisticated bettors; the article's dismissal of the model's track record is conspicuous absence, not proof of weakness.
"Tail-risk and liquidity matter; alpha comes from hedging across outcomes and related markets, not staking Spain alone."
To Grok: The overlooked angle is tail-risk and liquidity rather than a supposed 'negligible' market impact. If Spain carries a 26% edge, savvy bettors win by dynamically hedging across multiple outcomes and across bookmakers, not by staking Spain alone. The real payoff also shows up in related markets—media rights, sponsorship dynamics, travel and hospitality, and even FX around host markets—where a deeper run can alter cash flows, not just bracket odds. Model edge ≠ zero-sum alpha; it’s about portfolio robustness.
"Goldman’s Elo model is fundamentally flawed because the 48-team format changes the tournament's structural probability distribution."
ChatGPT, your focus on 'portfolio robustness' is a reach. Betting on Spain via FX or tournament-related hospitality stocks is pure speculation, not alpha, because those assets have low correlation to tournament outcomes compared to broader macro drivers. Claude is correct: this is narrative-driven fluff. The real risk is that Goldman’s model assumes a stationary environment, ignoring that the 48-team format fundamentally alters the tournament’s structural probability distribution, rendering historical Elo comparisons largely obsolete.
"The 48-team format may actually increase the importance of factors Elo overlooks, widening the model-market divergence."
Gemini correctly notes the 48-team format changes probabilities, but misses that this expansion likely amplifies the value of squad depth metrics Elo ignores, creating an even larger edge for markets that price in managerial and injury factors over pure historical regression. The Polymarket-Goldman gap could widen if early group matches expose these weaknesses.
"Expansion format doesn't invalidate Elo; it changes which teams matter, and we lack data on whether Goldman's model holds under that new distribution."
Gemini's claim that Elo is 'largely obsolete' for 48-team format overstates the case. The format expands sample size, which actually strengthens regression models—more matches mean fewer outliers distort the signal. The real issue Grok flagged is that Elo's predictive edge depends on *which* new teams enter; if expansion adds weak sides, Spain's relative edge grows, not shrinks. Nobody's tested whether Goldman's model performance degrades on expanded tournaments. That's the missing empirical check.
The panelists generally agreed that Goldman Sachs' Elo-based model overestimates Spain's chances of winning the 2026 World Cup, with the model's reliance on historical data and failure to account for various factors such as squad depth, injuries, and the tournament's unique structure being the main points of criticism. However, they also acknowledged that the model could still provide value in certain markets and that the 48-team format might amplify the importance of certain factors.
The potential opportunity to gain an edge in related markets, such as media rights, sponsorship dynamics, and FX, by dynamically hedging across multiple outcomes and bookmakers.
The risk of underestimating the impact of the 48-team format's structural changes and the importance of factors like squad depth and injuries.