Polling firm Wedgewood admitted on Oct. 20 that it fabricated most of its election surveys, and VoteHub, one of the aggregators that carried its results, is also the data source for recurring Kalshi contracts.
Kalshi’s September 2025 regulatory filing names VoteHub as the source agency for those contracts and states that “revisions to the Underlying made after Expiration will not be accounted for in determining the Expiration Value.”
The two facts raise a question about what happens to a settled contract when the polling number behind it is later withdrawn or found false.
Researchers at FiftyPlusOne requested raw data from Wedgewood’s Mississippi Senate survey and found inconsistencies in respondent identifiers and transaction records. They commissioned a comparison survey through the same provider, Pollfish, and its records looked materially different.
Wedgewood then admitted the fabrication while describing one Nebraska poll as legitimate. VoteHub’s head of data science, Zachary Donnini, said he spoke with a Wedgewood representative before including the firm’s polling, and that VoteHub graded Wedgewood a D and incorporated its data anyway.
The Guardian reported that VoteHub was among the polling databases that carried the results.
What Kalshi’s filing says
The Sept. 12, 2025 certification for Kalshi’s VOTEHUBAPPROVE contracts designates VoteHub’s time-adjusted polling average as the underlying and fixes the settlement value at expiration.
A fixed observation time is ordinary in event contracts, because traders need a definitive reference. The design also ties settlement to the reliability of the source data available at that moment.
A separate provision says that if the relevant VoteHub average is missing, the contract uses the most recent eligible average before the specified date. Treatment of a later retraction under that fallback falls to the rulebook.
Kalshi’s Oct. 20 generic congressional ballot contract, KXVHGCB-26OCT20, resolves Yes when VoteHub’s time-weighted Democratic-minus-Republican margin matches a specified percentage at the observation time.
It covers what a VoteHub measure will read at a precise moment and sets the election winner aside. Traders hold exposure to a third party’s methodology, publication schedule, and correction procedures.
The 2025 filing describes a reference reading at the first available 10 a.m. Eastern reading past the specified date, and the Oct. 20 market page specifies 5:05 p.m. Eastern the same day, so Kalshi’s current-series rulebook is the document to obtain.
For now, the filing establishes that the revision exclusion exists, and its application to the October series awaits confirmation.
Information moves in both directions
VoteHub’s 35-page methodology says its Complete forecasting model combines polling, fundamentals, expert assessments, and Kalshi prices.
Market weight varies by race, with more influence where contests are competitive or polling is sparse, and it has a cap. A separate Legacy model runs on polling and fundamentals, and the document acknowledges that feeding market prices into forecasts consumed by traders can create feedback loops.
The settlement reference is a polling average, and the forecast is a different data product. The methodology places market prices in the forecast and polling in the average, which keeps the settling average on its own inputs.
The practical issue for traders is whether a polling observation, a model forecast and a tradable probability stay clearly separated when all three appear in the same information environment.
What a correction would mean on Kalshi
A contract requiring 8% resolves Yes when the polling average reads 8% at the observation time. If the source later corrects the reading to 7.8%, the exclusion of post-expiration revisions keeps the original settlement, assuming the rule operates as written.
Settlement certainty is the benefit, and a payout built on a number later shown to be fabricated is the cost. The tension is sharpest for a fabrication than for an ordinary methodology update.
Whether this happened remains undocumented in the public record. Wedgewood’s identified fabricated surveys covered state races, and the example contract tracks a national average, so any effect of Wedgewood data on that average or on a payout remains unshown.
An archived page for the Oct. 20 contract displayed inconsistent volume figures across snapshots, so the article omits them.
In August, the Associated Press reported that a Kalshi contract tied to a candidate moved about two cents within 15 minutes of a fabricated Los Angeles mayoral poll from Median Strategies.
A cluster of Polymarket accounts traded in the same direction shortly afterward, and AP’s reporting left the trades’ ownership and any profits undetermined. That case involves a contract on an election outcome and sits apart from settlement against a polling average.
Where this goes from here
If Kalshi confirms which rulebook governs the October series, archives settlement inputs with timestamps and describes a dispute path for fabricated source data, the VoteHub contracts keep their certainty alongside a documented safeguard.
VoteHub could add value by publishing which Wedgewood surveys entered its averages and when they were removed.
If a contract settles on a reading later withdrawn and the exclusion rule blocks any adjustment, the first payout built on discredited data would test both exchange and aggregator.
Election contracts will multiply toward Nov. 3, and each one rests on third-party data screened before settlement becomes final.
Three answers from Kalshi would close the open points: which rulebook governs the October series, whether any contract has settled on a later-corrected figure, and whether a settlement can be reopened once fabrication is confirmed.