Brazil 2-1 snatches a late win over Japan, with Germany eliminated: How did prediction market odds reflect the match situation in advance?

On June 30, 2026, the Round of 32 of the USA-Canada-Mexico World Cup witnessed two moments that would go down in tournament history. At NRG Stadium in Houston, Brazil relied on Martinelli's 96th-minute stoppage-time winner to stage a 2-1 comeback against Japan and narrowly advance. Almost simultaneously in Boston, Germany lost 4-5 on penalties to Paraguay, failing to reach the Round of 16 for the third consecutive World Cup.

Two matches, two kinds of upsets—one where the favorite barely escaped in the dying moments, another where the favorite crumbled at the penalty spot. But for observers focused on crypto prediction markets, these games meant much more than just sports. They formed a natural laboratory for testing the pricing efficiency of prediction markets: To what extent did Polymarket's pre-match odds already reflect these outcomes?

What the pre-match probability distribution reveals about market consensus

Before Brazil vs. Japan kicked off, Polymarket's single-match market attracted over $3.14 million in trading volume, making it one of the most active matchups in the Round of 32. The market's pricing structure showed clear layers: Brazil to win in regular time was priced at 56 to 58 cents per share, implying a 56% to 58% probability; a draw was priced at 25 to 26 cents, implying 25% to 26%; Japan to win was priced at 18.5 to 19 cents, implying about 19%.

This set of data already conveyed important information. When other title favorites like France or England face lower-ranked opponents, their win probability usually exceeds 70 cents. Brazil, as a five-time World Cup champion, was only priced at a 58% regular-time win probability against Japan—this itself was a signal: the market did not view Brazil as an overwhelming favorite but regarded Japan as a genuine threat.

In the advancement market covering extra time and penalties, Brazil's probability to advance was about 76 cents, Japan's about 24 cents. The approximately 18-percentage-point gap between the 58% regular-time win rate and the 76% advancement probability revealed a deeper market judgment: if Brazil could not settle the match within 90 minutes, they were still more likely to prevail in extra time or penalties.

The pre-match pricing for Germany vs. Paraguay showed a different structure. Polymarket data indicated Germany's win probability at about 62%, Paraguay's upset chance at about 18%, and a draw at about 20%. The depth of the Germany -1.5 handicap market reflected the market's acknowledgment of Paraguay's defensive resilience. From the odds structure, Germany at 1.37, draw at 5.26, and Paraguay at 11.11 demonstrated that while the market favored Germany, it gave notable weight to the possibility that Paraguay could drag the game into extra time or penalties.

Dynamic coupling between market pricing and match progression

The Brazil vs. Japan match almost perfectly validated the script implied by the market pricing. Japan took the lead in the 29th minute through Kaiho Sano, which closely matched the pre-match assessment on Polymarket that "Japan to score over 0.5 goals was priced at 63 cents"—the market believed there was a 63% chance Japan would score at least one goal. Brazil equalized in the 56th minute through a Casemiro header, and the match then entered a prolonged stalemate.

With six minutes of second-half stoppage time, Martinelli completed the comeback in the 96th minute. Brazil advanced 2-1—exactly the "narrow win" script sketched by the market beforehand. Brazil's -1.5 handicap was only 31 cents, indicating that from the start, the market did not expect Brazil to win big. The deviation between the final result and the pre-match probability distribution was mainly that the 18-19 percentage point "Japan upset" probability did not materialize, rather than a misjudgment of the basic match structure.

The Germany vs. Paraguay match was a different narrative. Germany dominated possession in regular time but failed to convert dominance into goals. Paraguay broke the deadlock first, then Germany equalized, ending 1-1 after 120 minutes. In the penalty shootout, Germany's Havertz, Woltemade, and Jonathan Tah all missed, and Paraguay won 4-3 on penalties.

The key to this match: The market's pricing of a "draw" already contained the path to an upset. A 20% probability for a draw, combined with Paraguay's 18% upset probability, meant the market believed there was nearly a 40% chance Germany could not finish the game in regular time. When it finally went to penalties, the balance of victory was no longer about strength but about psychology and luck.

Information efficiency in the odds structure: which signals were priced

The core value proposition of prediction markets lies in information aggregation. When a large number of traders use real money to express their probability estimates of an event, market prices should theoretically reflect the weighted consensus of all available information.

From the Brazil vs. Japan pricing, the market conveyed several key signals: Japan has the ability to score (Japan over 0.5 goals priced at 63 cents), both teams are likely to score (both teams to score at 57%), Brazil will find it hard to win big (Brazil -1.5 only at 31 cents). These signals were verified one by one during the match, indicating that the market accurately judged the fundamentals.

However, from the Germany vs. Paraguay pricing, while the market correctly identified the risk that Germany would not win easily (the 20% weight for a draw was not low), it failed to fully price in Germany's vulnerability in the penalty shootout. Germany previously held a perfect World Cup penalty shootout record, and this historical data may have been over-weighted in the pricing model, overlooking the team's psychological fluctuations and offensive efficiency issues in crucial games in this tournament.

It is worth noting that the pricing efficiency of prediction markets is not reflected in "predicting accurately" but in "continuous correction." As the match progresses, market prices adjust in real-time: after Japan scored, Brazil's win probability might have briefly dropped; after Brazil equalized, the draw probability rose; during stoppage time, the probability of a late winner was repriced. This dynamic pricing mechanism is the core difference between prediction markets and static betting odds.

The logical connection between the scale expansion of crypto prediction markets and pricing depth

The 2026 World Cup has become a milestone in the history of crypto prediction markets. In the first quarter of 2026, on-chain prediction market trading volume reached $36 billion, surpassing traditional on-chain casino gambling for the first time. In the third week of June, on-chain prediction markets recorded a single-week trading volume of $10.8 billion for the first time, setting a historic record. During the World Cup's opening phase alone, daily trading volume exceeded $5.5 billion.

The expansion of scale directly impacted pricing depth. Polymarket's World Cup winner contract has seen trading volume exceed $3 billion, and the single-match market for Brazil vs. Japan with over $3.14 million in volume provided ample liquidity support for pricing. Bernstein estimates that this World Cup will bring over $3 billion in incremental trading volume to prediction markets.

There is a positive correlation between pricing depth and information efficiency. The larger the trading volume and the more diverse the participants, the harder it is for a single capital to manipulate market prices, and the more effectively they reflect the weighted consensus of dispersed information. From this perspective, the explosive growth in prediction market trading volume during the World Cup itself enhances market pricing efficiency.

Looking at the boundaries of risk pricing in prediction markets from two upsets

The matches between Brazil vs. Japan and Germany vs. Paraguay perfectly illustrate two different aspects of risk pricing in prediction markets.

In the Brazil vs. Japan match, the market successfully identified the core contradiction of "Brazil may win but won't win easily" and disassembled this judgment into tradable sub-dimensions through multiple layers (regular-time win probability, advancement probability, handicap, over/under). Although the final result ended with a Brazil stoppage-time winner, the match progression closely matched market expectations. In this case, prediction markets demonstrated not "predictive ability" but "structural analysis ability"—they accurately identified the most likely way the match would unfold.

In the Germany vs. Paraguay match, the market's performance was more complex. The 62% regular-time win probability pricing was not inherently wrong—Germany did dominate the game in regular time. However, the market failed to fully price in Germany's uncertainty in the penalty shootout. A penalty shootout is essentially a high-variance event, where the correlation between result and team strength is much lower than in regular play. When dealing with high-variance, low-predictability events, the pricing efficiency of prediction markets is naturally limited—this is not a flaw in the market mechanism but a reflection of the nature of the event itself.

These two matches together illustrate a core proposition: Prediction markets' pricing efficiency excels on "analyzable structural factors" but has natural boundaries on "high-randomness factors."

Reassessing the value of prediction markets as information aggregation tools

As the World Cup progresses, prediction markets are no longer niche experiments within the crypto industry. During the 2024 US presidential election, Polymarket's predictions of the outcome were generally more accurate than traditional polls. This event propelled prediction markets from a previously niche crypto product into the mainstream.

Sports provide a natural setting for prediction markets to directly compete with traditional betting. Unlike traditional sports betting, where prices are set by bookmakers adjusting for risk exposure, prediction market prices are determined collectively by market participants. This decentralized pricing mechanism makes prediction market prices better reflect "the wisdom of the crowd" rather than "the bookmaker's judgment."

From the two matches of Brazil vs. Japan and Germany vs. Paraguay, prediction market pricing did capture subtle information that traditional odds might overlook—such as the market's recognition of Japan's scoring ability and respect for Paraguay's defensive resilience. These signals might have been obscured by "favorite premium" in traditional betting markets but were clearly presented in the multi-dimensional structure of prediction markets.

The true value of prediction markets is not "accurately predicting outcomes" but "transparently presenting consensus." They aggregate judgments from traders dispersed around the world with different information into a readable, tradable, and traceable price signal. This signal itself is a manifestation of information efficiency.

Summary

Brazil's 2-1 stoppage-time victory over Japan and Germany's 4-5 penalty loss to Paraguay—two Round of 32 matches that interpreted the meaning of "upset" in completely different ways. From the perspective of prediction markets, these two matches respectively validated the strengths and boundaries of market pricing efficiency: when facing analyzable structural factors (Brazil would struggle to win big, Japan had scoring ability), market prices provided precise signals; when facing high-randomness factors (the outcome of a penalty shootout), market pricing was limited by the event's inherent unpredictability.

The 2026 World Cup has pushed the trading volume of crypto prediction markets to unprecedented heights—$10.8 billion in a single week, daily trading volume exceeding $5.5 billion. With continuous liquidity injection and the expanding participant base, the pricing efficiency of prediction markets continues to evolve. The World Cup is not only a feast of football but also the ultimate testing ground for prediction markets as information aggregation tools.

FAQ

Q: What was the probability distribution given by Polymarket before Brazil vs. Japan?

A: As of June 29, 2026, Polymarket data showed Brazil's regular-time win probability at 56% to 58%, a draw at 25% to 26%, and Japan's win probability at about 19%. In the advancement market, Brazil's probability to advance was about 76%, Japan's about 24%.

Q: What pricing did the prediction market give before Germany vs. Paraguay?

A: Polymarket data showed Germany's win probability at about 62%, Paraguay's upset probability at about 18%, and a draw at about 20%. The odds structure was Germany at 1.37, draw at 5.26, and Paraguay at 11.11.

Q: How well does prediction market pricing efficiency perform in sports events?

A: Prediction markets show high pricing efficiency on "analyzable structural factors," such as accurately capturing signals like Brazil would struggle to win big and Japan had scoring ability. However, on "high-randomness factors" like the outcome of a penalty shootout, pricing efficiency is limited by the event's inherent unpredictability.

Q: What level of trading volume did prediction markets reach during the 2026 World Cup?

A: In the third week of June 2026, on-chain prediction markets recorded a single-week trading volume of $10.8 billion for the first time, a historic record. During the World Cup's opening phase, daily trading volume exceeded $5.5 billion. Polymarket's World Cup winner contract has seen trading volume exceed $3 billion.

Q: What is the core difference between prediction markets and traditional sports betting?

A: Prediction market prices are determined collectively by market participants, reflecting the weighted consensus of dispersed information; traditional sports betting odds are set by bookmakers adjusting for risk exposure. The decentralized pricing mechanism of prediction markets allows them to better embody "the wisdom of the crowd."