Same match, same brand, two settlement prices.

Two prediction-market products carry the Polymarket brand. Polymarket International is the older, offshore platform. Polymarket US operates through a separate corporate structure designated by the CFTC as a contract market — QCX LLC d/b/a Polymarket US, status “Designated” since 2025-07-09, per the Commission’s own filing record and its signed Amended Order of Designation.

Both listed moneyline markets on the same two tennis matches. Both matches ended in a walkover — a player withdrew before the first point was played, and the opponent advanced. Each venue settled under a rule it already published for exactly that situation. The rules are different, and so were the payouts.

That the two platforms publish different walkover rules is not new. Each states its rule in its own documentation, and the difference has been described in general terms before. What this page adds is narrower: two real, independently verifiable matches where those rules produced different settlement values on the same underlying event.

Matteo Arnaldi was due to play Flavio Cobolli in the men’s singles semifinal on 2026-06-05. He withdrew before the match began, citing a viral illness, and Cobolli advanced to face Alexander Zverev in the final.

Polymarket International settled both sides at 50¢. Polymarket US settled Arnaldi at 33¢ and Cobolli at 67¢. In dollar terms, for two round-number positions:

These are settlement values, not trading profits. We have not reconstructed what any trader paid to hold these contracts, and we cannot say anyone made or lost the difference. What we can say is that the two venues assigned different dollar values to comparable positions on the same match.

The Polymarket US market on this match traded actively, and most of that trading happened after Arnaldi’s withdrawal was already public. A separate case file reconstructs that market’s full tape and owns every volume figure for it; this page states none.

Elina Svitolina withdrew from her third-round match against Xiyu Wang at the Cincinnati Open, a WTA 1000 event, on 2026-08-18, citing a right ankle injury. Wang advanced by walkover into the round of 16.

Polymarket International again settled both sides at 50¢. Polymarket US settled Wang at 26¢ and Svitolina at 74¢ — close to where the market had priced Svitolina as the favorite before she pulled out.

Why this page publishes no volume figure for the Cincinnati market. Polymarket US publishes volume figures for this market that fail the venue’s own published definition of what those figures mean, by about two orders of magnitude. We checked them against the market’s own trading record and established which figure is wrong, but a number that requires us to correct the venue’s own published field before quoting it is not one we will put on a page. The settlement prices are unaffected — they come from three separate first-party records that agree.

Polymarket International’s rule for the markets we examined fixes the outcome in advance:

Polymarket US’s two contracts do not fix a number. They point at the market instead, and they do it in two different ways. The Roland Garros contract, created in June, says:

The Cincinnati contract, created in August, says:

These are genuinely different template wordings, not one rule paraphrased — “last-traded prices” in the first, “fair market price” in the second. Neither names a time. What they share is the direction of the result: where International’s flat rule erases whatever the market thought, both US contracts preserve it.

On both matches the market had a clear favorite before the withdrawal, and on both, that is roughly where the US settlement landed. International paid both sides the same amount regardless.

There is a second divergence, and it sits inside Polymarket US rather than between the two platforms.

The Roland Garros contract’s own text settles a walkover at “last-traded prices”, with no time anchor. Polymarket US’s Sports FAQ, archived and confirmed live since at least March 2026 — three months before that match — states a different rule for the identical scenario: that anything happening before the first serve is a pre-event scenario settling at last fair market prices as of the official announcement, except for ITF matches, which resolve at $0.50 per contract. A Grand Slam ATP semifinal is not within the ITF carve-out.

So two contemporaneous first-party Polymarket US documents state different rules for the same situation: one anchored to an announcement, one anchored to nothing. Neither reading reproduces the settlement from the last pre-announcement trade, which printed at 32¢.

Which document governs is settled; which was applied is not. Polymarket US’s rulebook says a contract’s own terms control where they conflict with the general rules, so the contract text governs as a matter of hierarchy. That does not tell us which rule produced the number, and the public record does not show it. We searched the exchange’s notice archive for any record of a formal outcome review on this contract and found none — which is not evidence that none occurred.

• We are not saying either venue calculated its settlement incorrectly. Each did what its own published rules say it would do.
• We are not alleging manipulation, insider trading, or that anyone exploited the difference.
• We are not claiming any trader made or lost the settlement-value difference shown above. We have not reconstructed entry prices or full position economics.
• We are not claiming that every cross-listed tennis walkover settles differently on the two platforms. We verified two, which is not a census.
• We are not claiming the rule difference was secret. Both venues publish their rules.
• We are not saying Polymarket International and Polymarket US are the same exchange or the same legal entity. They are not. That does not change what a customer sees: two products under one brand, on one match, settling to different values.

The questions this leaves are for Polymarket:

• Why do the two platforms use different settlement methods for a comparable tennis walkover?
• What is the reasoning behind International’s flat 50¢ rule against the US platform’s market-price-based rule?
• How were the 33¢/67¢ and 26¢/74¢ settlements calculated?
• Where a US contract’s own text and the Sports FAQ state different rules for the same scenario, which is applied in practice?
• Does Polymarket track how often comparable cross-listed markets settle to different values across its two platforms?
• Are there plans to align settlement methodology between the two?

Sources. Settlement prices on the US side come from Polymarket US's public market API, cross-checked against two further first-party records. Settlement prices on the International side come from Polymarket International's public API, fetched one market at a time. Rule text is quoted from each contract’s own published terms, not from marketing copy or help-centre summaries.

Why we are confident each price belongs to the player we say it does. Pairing a settlement price with the wrong player would invert this entire page, so no price here rests on a single record. Each one was confirmed against three independent first-party records from the same venue, and all three agree in both matches. Where a venue’s own records could be read more than one way, we used the reading the venue itself labels player by player.

Our own screening produced false positives, and we kept them. An early wide pass over closed markets threw up a large number of apparent 50¢/50¢ settlements that turned out to be ordinary completed matches, and a separate automated read of a tournament draw reported a walkover in a match that had actually been played. Neither survived contact with the underlying records. Every settlement on this page was confirmed market by market before it was published, and both false positives are kept in our evidence set as a standing caution.

A third candidate we ruled out. A January 2026 match was checked as a possible third case and rejected: Polymarket US had not yet listed tennis at that date, confirmed against its full market listing for that week rather than assumed.

Limits. Two matches. This page establishes that the divergence is real and realized, not how often it happens. We did not census cross-listed walkovers, and no rate should be inferred from a two-for-two result.