Why most mention market buyers lose money

A mention market is a popular prediction market contract that pays if a named person says a specific word or phrase during a specific event. Platforms such as Kalshi and Polymarket list these contracts across several recurring event categories:

• Presidential speeches and debates
• Federal Reserve press conferences
• Corporate earnings calls
• Live sports broadcasts
• Late night television shows

On a typical event board, exchanges list dozens of individual word contracts spanning prices from a few cents to near-certainty.¹ Buyers frequently take Yes positions on these contracts under the assumption that public speakers follow predictable speech patterns or repeatable scripts,² or are scrupulous.

These assumptions are fundamentally flawed.

Across more than 35,000 completed Kalshi mention markets analyzed by the Wall Street Journal, traders opening Yes positions at initial listed prices lost an average of 11 percent.³⁴ These losses stem directly from structural realities: Yes contracts enter the market systematically overpriced, settlement hinges on legal rulebooks rather than viewer perception, and outcomes remain vulnerable to speaker and insider manipulation.

Yes contracts consistently lose money.

Yes contracts lose money consistently across major event categories because entry prices systematically overstate true payout frequencies. Across a Wall Street Journal analysis of more than 35,000 completed Kalshi mention markets, buyers opening Yes positions at initial listed prices lost an average of 11 percent.⁵⁶

This loss stems from structural overvaluation across the probability spectrum. Retail buyers routinely treat low-priced Yes contracts like lottery tickets, bidding prices up far beyond their statistical chance of occurring. Even contracts listed at an implied 50 percent probability paid out only around 40 percent of the time in practice.⁷

Speed and order book friction compound these pricing losses. Mention markets update within seconds once a target word is delivered, meaning late pricing reflects an already settled result rather than a forecast or predictive edge.

Furthermore, thin trading volume across secondary word options creates wide gaps between buy and sell prices. A buyer paying 50 cents for a contract may face a top sell offer of only 40 cents, instantly surrendering a portion of their capital to execution friction before the event even begins.

Contract rules decide which words count.

Contract rules decide which words count by enforcing legal definitions over spoken audio or viewer perception (or hope). When a trader hears a phrase during a stream, they often assume the contract resolves to Yes. However, exchange settlement depends entirely on text criteria written into the rulebook.

Kalshi standard terms require the exact listed word, plural, or possessive form.⁸ Rules explicitly exclude grammatical or tense inflections and disqualify statements picked up on hot mics.⁹ A trader dispute log shows how legal text overrides broadcast audio:

• Spoken abbreviation: "Warner Brothers" spoken on an earnings call resolved the "Warner Bros" strike to No because the abbreviated text was required.¹⁰
• Verb inflection: Announcers saying "turfs" as a verb resolved the "Turf" strike to No.¹¹
• Unapproved verification: Listed words spoken at a political event resolved to No because mainstream press were barred from the venue, leaving designated primary sources without coverage.¹²

Resolution authority rests on platform-specific terms rather than general consensus. Kalshi relies on video and official transcripts, while Polymarket uses decentralized reporting backed by a $750 bond and token-holder voting.¹³ Because exchanges settle strictly according to full contract documentation, minor discrepancies between broadcast audio and transcript logs often cause Yes positions to resolve to No.¹⁴¹⁵

Cancelled events pay zero to Yes buyers.

Cancelled events pay zero to buyers because any event cancellation or venue change automatically resolves every word contract on the board to No. When a speech, press conference, or broadcast is cancelled, delayed past deadline, or moved to an unapproved room, exchanges do not issue refunds or hold funds in limbo. Instead, the entire board wipes out Yes holders.

Kalshi rulebooks enforce this through a master clause: if an event fails to meet listed conditions, a dedicated contract titled "Event does not qualify" resolves to Yes, which forces every listed word option on that board to resolve to No.¹⁶ On major address boards, this disqualification contract typically trades around 3 cents, pricing the baseline risk that a sudden schedule disruption wipes out the board.¹⁷

Venue parameters are applied with absolute strictness. For example remarks delivered during hallway interviews, side press conferences, or informal side meetings do not count toward settlement if they occur outside the official event boundaries.¹⁸

The harsh reality of mention markets is that traders holding Yes positions bear total downside risk from event disruptions without any refund mechanism when a broadcast is altered.

Insiders and speakers can control outcomes.

Insiders and speakers control outcomes by using advance (or confidential) information or directly manipulating speech scripts during live broadcasts. While retail traders attempt to predict word choices from historical speeches, participants with early access or direct microphone control hold total operational control over contract resolution.

Unsurprisingly, inside information creates severe market asymmetry. A notable 2026 CFTC enforcement action against a former White House teleprompter operator documented how insider access to speech drafts was used to systematically trade presidential mention contracts on Kalshi prior to delivery, leading regulators to order over $172,000 in civil penalties and profit disgorgements alongside a three-year trading ban.¹⁹²⁰

Speakers can also manipulate contract outcomes directly from the live microphone. On a corporate earnings call, Coinbase's chief executive concluded the broadcast by explicitly reciting a list of target words: "I just want to add here the words Bitcoin, Ethereum, blockchain, staking and web3 to make sure we get those in before the end of the call".²¹

Trade the structural realities of mention markets over trivialities.

Mention markets are risky, easily manipulated, and among the most prone to settlement disputes across prediction platforms. Anyone participating in them must trade with eyes wide open, size positions strictly, and accept the severe legal and structural friction favoring contract sellers.

Ultimately, guessing verbatim speech snippets represents low conviction media noise. Sustainable edge in prediction markets comes from applying structured research and a disciplined process to meaningful, real world outcomes that can be predicted with genuine accuracy.