By Frank, PANews
The 2026 FIFA World Cup in the U.S., Canada and Mexico has ended, and so has one of the biggest prediction-market trading runs tied to a global sporting event. Data compiled by PA Beacon and cited by PANews shows that after all 104 matches were completed, large pre-match traders on Polymarket had put $474.04 million to work before kickoff. Of that total, $298.49 million landed on the right side, lifting the dollar-weighted hit rate to 62.97%.
Using a buy-and-hold-to-resolution method, those positions would have returned about $502.93 million, for an estimated net profit of $28.8858 million and an ROI of 6.09%.
That aggregate result looks clean. The account-level picture does not. A small number of wallets captured a large share of the gains with one or two oversized positions, while other accounts spread bets across most of the tournament, posted higher hit rates, and still earned far less. The profit leaderboard and the consistency leaderboard were not the same thing.
How the sample was built
The study covered all 104 World Cup matches and 312 win, loss and draw markets. Trade data came from the Polymarket Data API. The filter only included pre-match BUY transactions worth at least $5,000. Multiple buys from the same wallet in the same match, same market and same direction were merged into one aggregated position.
Profit and loss was estimated on a hold-to-resolution basis. The calculation did not include mid-match sales, in-play trading, hedges in other markets, or fees, so the figures do not represent each wallet’s actual realized final P&L.
PA Beacon had previously looked at the first 20 matches on June 17, 2026. At that stage, pre-match buying totaled $89.5457 million, the dollar-weighted hit rate was only 48.5%, and the group as a whole was down an estimated $1.7594 million, with ROI at -2.0%.
One month later, the full-tournament picture had flipped. Large traders finished the event with a positive return.
Profits turned positive, but they were concentrated
Among 1,856 wallets that placed qualifying pre-match buys, 1,003 were profitable under the study’s methodology and 853 lost money. That puts the share of profitable wallets at 54.04%.
Winning wallets made a combined $99.6338 million, while losing wallets gave up $70.7480 million. Netting those two sides leaves the total estimated profit at $28.8858 million.
The concentration was striking. The top five profitable accounts earned $37.7064 million, or 37.85% of the gross profit booked by all winning wallets. The top 10 earned $55.1869 million, equal to 55.39%.
So the market’s overall 6.09% expected gain did not mean most large traders had found a stable edge. A handful of concentrated winners pulled the average higher.
Match-by-match dispersion was just as sharp. Belgium vs. Egypt drew $12.3905 million in pre-match buying, but only $667,000 was on the correct side, for a dollar-weighted hit rate of 5.4%. Belgium vs. Senegal drew a similar $12.0438 million, yet $10.5326 million of that total was correct, producing a hit rate of 87.5%.
Similar amounts of money produced completely different outcomes. Size alone did not make a position right.
A few oversized bets created the most visible winners
The clearest example was the account named “mintblade.” It participated in only two matches and three aggregated positions, deployed $7.2889 million, and generated an estimated $8.2535 million in profit. Its ROI reached 113.23%, and its dollar-weighted hit rate was 100%.
One match, Iran vs. New Zealand, accounted for $6.7738 million of that profit, or 82.1% of the wallet’s estimated World Cup gain.
The setup was simple. “mintblade” spent $6.4705 million buying “Iran not to win” at an average price of about $0.49. The match ended 2-2, so the position resolved at $1 and returned an estimated $13.2443 million. The account then bought “Uruguay not to win” and a small draw position in Saudi Arabia vs. Uruguay, adding about $1.4797 million more in profit across the two trades.
Two correct calls were enough to place the wallet among the most profitable accounts of the tournament.
The wallet behind “GRIMDRIP” was even more concentrated. It traded only one match, Czech Republic vs. South Africa, and bought two related directions: “Czech Republic not to win” and “match draw.” The final score was 1-1, both positions resolved in its favor, and $5.8490 million in capital produced an estimated $7.4497 million in profit. ROI came in at 127.37%.
“DEEDDIT” represented a different version of concentration. That account covered eight matches and nine aggregated positions, deployed $20.9493 million, and earned an estimated $8.0636 million. In the round of 32, Belgium beat Senegal 3-2, and the account’s $7.1610 million bet on Belgium to win generated an estimated $7.7495 million in profit. In the semifinal, France lost 0-2 to Spain, and a “France not to win” position added another $3.4259 million.
Still, “DEEDDIT” was not consistently right. It bought the wrong draw side in matches including Mexico vs. Ecuador and Switzerland vs. Colombia, with major losses exceeding $4.3 million. Its biggest winning position alone equaled 96.1% of its net World Cup profit, showing how heavily the result depended on Belgium vs. Senegal.
The accounts “sparklingwater123” and “endlessFate” had similar profiles. The former covered only two matches, put in $8.5005 million, and earned an estimated $7.7469 million. The latter covered five matches, deployed $11.4454 million, and earned an estimated $6.1929 million. PANews said these wallets repeatedly bought “not to win” or draw outcomes when prices on direct wins for popular teams looked too expensive, though they did not simply fade favorites across the board. Their view was expressed in a small number of concentrated matches.
High participation did not translate into outsized profit
If the ranking is sorted by frequency rather than profit, “swisstony” stands out. The account covered 97 matches and 267 aggregated positions, making it one of the broadest participants in the sample. It put $15.9912 million into pre-match buys, recorded a dollar-weighted hit rate of 79.29%, and earned an estimated $1.2482 million, with ROI at 7.81%.
It called far more matches correctly than “mintblade” or “GRIMDRIP,” yet its profit was only about 15% of “mintblade’s.” Position structure explains most of the gap. “swisstony’s” biggest estimated gain on a single position was only $222,700, while its biggest loss was $305,800. Its largest winning position accounted for just 17.84% of its total World Cup net profit.
Rather than turning on one dramatic reversal, the account accumulated gains through smaller exposures across many matches.
“AV23IUa” and “Latina” also fit the broader-coverage winner profile. “AV23IUa” traded 46 matches and 46 positions, deployed $2.1245 million, and earned an estimated $906,000, for ROI of 42.65%. “Latina” covered 11 matches, spent $3.4293 million, and made an estimated $1.2920 million, with a dollar-weighted hit rate of 88.64%.
Another account worth noting was “zhqzhq.” Using the study’s standard of whether more than half of a wallet’s capital in a given match was on the correct side, it was right in all 14 matches it covered. Even so, $1.0182 million in capital translated into only $55,700 of estimated profit, for ROI of 5.47%.
That result showed how a high hit rate could still produce limited returns when positions were entered at points where the outcome was already heavily priced in.
Losing accounts offered the clearest contrast
The losing side made the trade-off even clearer. “LEEEROYJENKINS” posted an estimated $4.7976 million profit on Australia vs. Turkey, then took a heavy position on Belgium to beat Egypt. The match ended 1-1, and that single trade produced an estimated loss of $8.3943 million.
The earlier profit was erased, and the wallet finished the World Cup down an estimated $3.2464 million.
Other losing players were simply wrong over and over. “coldsway” deployed $13.7255 million across nine matches, posted a dollar-weighted hit rate of only 26.22%, and ended with an estimated loss of $7.3437 million, the biggest account-level loss in the sample. “FlickRaw” traded only two matches, got both wrong, and saw its entire $4.7983 million stake go to zero.
These cases showed that the wallets in this group still behaved like high-stakes gamblers. One oversized mistake was enough to wipe out many correct calls.
How whales priced the final contenders
As the tournament moved into its final stretch, France, England, Argentina and Spain became the four teams drawing the most attention in title discussions. Betting patterns around their knockout-stage direct win markets offered another way to read how large traders operated.
Spain: the most stable dollar signal
Spain was the cleanest case among the four teams. Across five knockout matches, the side with more money was correct every time. Support for a Spain win rose from $1.143 million in the round of 32 to $2.390 million by the quarterfinals, then fell to $879,000 in the semifinal against France, where the median entry cost was just $0.297.
Spain still won 2-0, making that match the one with the largest payoff room. Yet the money was not broadly distributed. The biggest account supplied 46.7% of the pro-Spain side, and the top five accounts contributed 73.5%.
Five straight correct outcomes reflected a few heavily committed wallets getting the direction right, not a uniformly correct market.
Argentina: support faded even as the team advanced
Argentina showed the opposite pattern. The team kept advancing and reached the final, but direct support money steadily declined. In the round of 32, support totaled $2.804 million and the median cost was $0.86. By the semifinal, those figures had fallen to $393,000 and $0.315.
Before the final, only $434,000 backed Argentina to win, while $1.518 million was placed against that outcome. Support accounted for just 22.2% of the total. In this case, capital had shifted to the “not to win” side in advance, and that side ended up being correct.
Again, the conclusion was heavily concentrated. The top five accounts on the anti-Argentina side contributed 92.5% of that capital, and the biggest one alone accounted for 44.4%.
France: the sharpest case of a few getting it right while the crowd got it wrong
France offered the clearest example of concentrated conviction. In the semifinal against Spain, $1.626 million backed France to win, while $7.194 million backed the other side. Large money was right in that match, but about $5.76 million of the anti-France capital came from a single wallet, equal to 80.1% of that side.
Then sentiment swung hard in the third-place match. A total of $2.9997 million backed France to win, while only $179,300 went the other way, putting support at 94.4%. France then lost 4-6 to England.
England: the team most often undervalued
England was the easiest of the four to underestimate. In the round of 16 against Mexico, $1.144 million backed England and $1.187 million backed England not to win. The dollar consensus leaned against England, but England advanced 3-2.
In the third-place match against France, only $159,000 backed England, while $1.032 million opposed that result. Support was just 13.4%, and the median entry price was only $0.210. England still won 6-4.
Even there, PANews argued that the result should not be read as a broad failure by smart money. In the third-place match, the top five anti-England accounts provided 98.3% of the capital on that side, and the biggest account alone accounted for 60.4%. A few concentrated wallets pulled the money-weighted consensus in the wrong direction.
The dataset showed how uneven “smart money” really was
Once the 40-day World Cup trading frenzy was over, the on-chain record pointed to a market where capital size did not equal foresight, and where money-weighted consensus could be skewed by a handful of wallets.
Some addresses made a tournament’s worth of profit from one upset or two carefully chosen matches. Others spread capital across nearly the whole event and produced steadier but much smaller returns. And on the losing side, one oversized error could wipe out a long string of correct calls.
The headline number was positive. The distribution behind it was anything but even.