The Deal That Never Happened
A prediction market said Ukraine signed a minerals deal it never signed, and the price was never the problem. This episode follows who gets to decide what happened, and why an audit that clears them can't actually clear them.
FactUnion VerifiedEvery fact on the episode's fact sheet was checked against its sources. Wrong facts were corrected. What this meansFacts checked 18 Oct 2026
Chapters (10)
About this episode
In March 2025 a Polymarket contract on a US–Ukraine minerals deal resolved "Yes", although no deal had been signed. Polymarket called it an unprecedented governance attack and refused refunds. Nadia and Miles rebuild the forty-eight-hour UMA vote that decided it. Then they follow the same voting bloc into the June 2026 Strategy bitcoin-sale dispute, the Wall Street Journal's finding that referees often held positions in the markets they were deciding, and Bitquery's October 2026 audit of 489 resolved votes. That audit found the nine largest wallets changed none of the outcomes.
The audit's clean result is real. It is also exactly what you would see if the system were captured, because UMA pays voters for matching the majority. The hosts test the explanations against each other: one whale, a bloc, a crowd guessing the crowd, plain bad rules. They look at Kalshi's in-house referee as a comparison and end with what evidence would actually settle it. Nothing in this episode is trading advice. Positions held, motives and coordination are described only as reported, and where nobody has said anything, the hosts say that too.
Hosted by Nadia and Miles, performed by synthetic voices. Supported by STELQ.
Facts and sources
Every fact used in this episode, with the sources it was checked against. Download as data
165 of 165 recorded lines matched the checked script.
A Polymarket market asking whether Ukraine would agree to a US minerals deal before April resolved ‘Yes’ in March 2025 even though no such deal had been signed.
CoinDesk reported that a contentious Ukraine bet led to a clash between the Polymarket and UMA communities.
CoinDesk reported that a wallet called BornTooLate.Eth had become a top-five UMA staker, holding more than 1.3 million UMA tokens, at the time of the Ukraine resolution.
According to a widely shared thread, the ‘Yes’ proposal on the Ukraine market was made on Saturday, March 22, 2025, when Polymarket staff were off duty, and the ‘Yes’ side relied on a February 27 statement by the US Treasury Secretary.
Contemporaneous accounts report that Polymarket posted a clarification saying it was too early for the market to resolve two minutes before UMA’s commit period ended, and then issued another clarification two minutes before the reveal period ended.
Show all 24 factsShow fewer facts
Polymarket called the Ukraine episode an ‘unprecedented’ governance attack by a UMA whale and declined to issue refunds.
- Polymarket says governance attack by UMA whale to hijack a bet’s resolution is ‘unprecedented’theblock.co
- Polymarket Reports 'Unprecedented' Governance Attack by UMA Whale on Bet Resolutioncoinmarketcap.com
- Whale Allegedly Manipulates $7 Million Bet in Polymarket Governance Attacknews.bitcoinprotocol.org
In UMA’s dispute system as used by Polymarket, the proposer and the disputer post equal bonds, typically $750, and the winner collects half of the opponent’s bond.
A disputed question goes to UMA’s Data Verification Mechanism for a 48-hour vote, with a 24-hour commit phase followed by a 24-hour reveal phase, and an outcome needs 65% of staked UMA to agree.
UMA voters who vote with the minority or miss a vote are slashed 0.1% of their stake per error or absence, and the slashed tokens are redistributed to the majority.
In August 2025 UMA updated its oracle so that resolution proposals for Polymarket markets were limited to whitelisted parties, a change known as the managed oracle or MOOV2.
An independent analysis by Fidonis found that dispute rates fell from 1.52% to 0.56% after the August 2025 whitelist (UMIP-189). It attributes the fall mostly to broader market trends and finds no statistically detectable reduction in contested outcomes when measured by volume.
In June 2026 UMA token holders voted 98.6% to resolve a Polymarket market on a Strategy bitcoin sale (a reported 32 BTC sale) to ‘No’, despite the transaction being confirmed.
Cointelegraph reported that blockchain wallet borntoolate.eth held 3.11 million UMA tokens in the Strategy dispute, followed by Kevin Chan with 1.53 million tokens.
Polymarket faced trader backlash in early June 2026 over the disputed Strategy bitcoin-sale market, with traders arguing the sale fell within the valid window. No public statement from Kevin Chan or borntoolate.eth responding to the controversy appears in our findings.
A Wall Street Journal analysis published May 18, 2026, using blockchain data linked to Polymarket accounts, found that in more than 300 disputes over the previous year, nearly 20% of those it reviewed, at least one UMA voter held a position in the market being decided.
- Nearly 1 in 5 Polymarket disputes show voter conflict of interestedgen.beta.edgen.tech
- Nearly 1 in 5 Polymarket disputes show voter conflict of interestedgen.dev.edgen.tech
- Wall Street Journal investigation: Polymarket arbitration system—anonymous token voters suspected of manipulating resultsgate.com
A UMA voter identified as ‘Scout’ acknowledged to the Journal regularly voting on resolutions while holding positions in the same contracts and defended the practice as necessary.
Bitquery’s October 9, 2026 investigation matched 470 Polymarket disputes to 489 resolved questions from UMA voting rounds between September 9 and October 7, 2026.
Bitquery found that the nine largest voting wallets held about 56.3% of average voting weight and acted as a near-unanimous bloc, but re-running the votes without them changed zero of the 489 outcomes.
Bitquery identified wallet 0xc816f934fab8179997a05e4b71f405b9e2f69eb8 as winning all 80 disputes it filed, mostly by challenging resolutions proposed too early.
Bitquery found four wallets that forfeited 500 USDC each across 46 disputes, about $23,000, in one night, all on markets for a Montevideo Wanderers vs. Albion football match.
Kalshi settles contracts internally under its rulebook, and a contested market goes to an Outcome Review Committee that makes a final determination within 24 hours.
Kalshi faces a class action, Risch v. KalshiEX LLC, over a reported $54 million market on the death of Iran’s Supreme Leader, in which traders argue that a ‘death carveout’ clause the exchange enforced was poorly disclosed.
According to a review of resolution failures, in January 2026 Kalshi mistakenly settled some NFL markets early, then after public backlash reversed course and paid affected traders from its own funds.
New Jersey has asked the Supreme Court to hear its fight with Kalshi (Flaherty v. KalshiEX, No. 26-299), supported by 39 states and the District of Columbia, and Kalshi has asked for until November 9, 2026 to respond.
Removed during checking (3)
On-chain attack documentation (OAK) describes about 5 million UMA tokens cast from three wallets, roughly 25% of the total vote, behind the 'Yes' resolution of the Ukraine market. The 1.3 million figure was one wallet's balance, not the total voting power used.
A UMA spokesperson described the Ukraine resolution as a series of unfortunate events, while some users argued that voters ignored Polymarket's clarification to avoid being slashed for voting with the minority.
Election-market volume on prediction platforms rose 136.8% from September to October 2026 as the midterms approached.
