The Polymarket contract asked a simple question: did Strategy (MSTR) sell any bitcoin by May 31? The company’s filing says it sold 32 BTC between May 26 and 31. The filing came out June 1. That gap has split bettors into a $79 million fight over whether a sale counts when it happens, or when it’s confirmed.

The dispute turns on a single ambiguity: the rules ask whether Strategy sold bitcoin “by 11:59 PM ET” on May 31, but they don’t say whether that means the sale must have occurred by then or been confirmed by then.

Strategy executed the trades between May 26 and 31 and dated the activity “as of May 31, 2026, 4:00 p.m. Eastern Time” — inside the window. But the 8-K disclosing them wasn’t filed until June 1, after the market closed. So the sale date falls before the deadline; the filing date falls after it. Which one governs is the whole fight.

The fight breaks into three camps, and it plays out in the language of UMA’s voting options. One says the market is event-based and should resolve “Yes” (P2), because Strategy’s own filing dates the sale inside the deadline.

Another says it is effectively announcement-based and must resolve “No” (P1), because nothing confirmed the sale before the market closed. A third invokes P4 — the “too early” vote, meant for proposals made before an event has occurred — arguing the rules were too vague to resolve until Strategy’s filing landed.

CoinDesk went through the dispute threads on both Polymarket and UMA’s Discord channels, and with the assistance of AI, summarized the arguments the different camps are making.

The ‘Yes’ case: the sale is what matters

This camp reads the market as event-based, pointing to rules that resolve “Yes” if the firm “sells any of its Bitcoin” by the deadline, with no requirement that the sale be announced by then.

Their evidence is Strategy’s own disclosure, which lists the 32 BTC as sold “during period May 26, 2026 to May 31, 2026” and presents the activity “as of May 31, 2026, 4:00 p.m. Eastern Time” — inside the window. Because the rules name “information from MSTR” as the primary resolution source, they argue, the source itself confirms the sale.

Several add that Strategy reports weekly, usually on Mondays, so a late-month sale could never be confirmed before a month-end deadline — making a “No” reading a bet on filing schedules rather than events.

The ‘No’ case: only what was knowable by the deadline counts

This camp treats the market as announcement-gated, citing past Polymarket markets that resolved using only information available within the timeframe. The 11:59 p.m. ET deadline, they argue, defines a closed window: new information can always arrive later, but it doesn’t reach back to change a settled outcome, and nothing had confirmed a sale when the answer was proposed June 1.

Some note that the “as of May 31” language the other side leans on only surfaced in that day’s filing. Underpinning it is an integrity argument — that if a dispute can hold a market open until favorable evidence appears, anyone could extend any deadline for the price of a bond.

The ‘too early’ case: the rules can’t resolve this yet

A smaller group argues the market was too poorly drafted to resolve cleanly either way, noting the rules require the sale to occur “on the date specified in the title” rather than “by” it, leaving no coherent timeframe.

With Strategy’s filing due imminently and named as the primary source, this camp contends the market should have stayed open until that disclosure published rather than being resolved on a deadline they consider malformed. The “No” camp’s reply: P4 doesn’t apply, because the sale itself predates the deadline — the proposal wasn’t early, the confirmation was just late.

Polymarket’s clarification, and the catch

Polymarket has since added context backing the “No” reading, stating that no information from MSTR, on-chain data, or credible reporting confirmed a sale within the timeframe and that “confirmation achieved outside of the market’s time frame does not qualify.” Traders priced it accordingly, with the May 31 contract collapsing from 81% “Yes” during the dispute to under 1%.

But Polymarket doesn’t cast the final vote — UMA’s token holders do, and the two have split before. In 2024, UMA voted that Barron Trump wasn’t involved in the DJT memecoin; Polymarket overruled the oracle and refunded “Yes” holders anyway. For now, the two appear aligned.

The sale everyone can see is trading at less than a penny.