A speech is a strange thing to price. It is not a vote, not a poll, not a scoreboard; it is one person at a lectern, choosing words in real time, and a room of people counting them. That is the peculiar exercise Polymarket has set up for Donald Trump's scheduled remarks in New York on 14 August 2026, where the contract does not ask what he will announce or whom he will attack. It asks which specific words come out of his mouth.
The question is narrow by design. And the resolution rule is unusually literal, which makes the contract worth reading carefully before treating any of the numbers on it as a forecast of anything larger.
Deposit $20, get $50 to trade
Welcome bonus on the world's largest prediction market.
18+ · New users only · Availability and terms vary by region · Trading involves risk of loss · BeGambleAware.org
What the contract is actually measuring
Strip the market down to its plumbing and it looks less like a political forecast and more like a transcription exercise with money attached. The Trump remarks word market on Polymarket offers a list of terms and phrases, and each one resolves Yes if Trump says it during the specified New York appearance on 14 August 2026, and No otherwise. Some entries are single words. Some are word-count thresholds, like "Crime 15+ times" or "Hell 7+ times". Some are names. One outcome, "-No Qualifying Event-", is the escape hatch for the scenario in which the appearance is cancelled or otherwise does not happen in a form that qualifies.
The list of listed terms includes Biden, Communist / Communism, Oil / Gas / Gasoline, Police Officer, Hottest, Stock Market, Crazy, Transgender, MAGA / Make America Great Again, Supreme Court, Kamala, Barack Hussein Obama, Zohran / Mamdani, "Think About It" and Bitcoin / Crypto / Cryptocurrency. The count-based rungs include Drug / Murderer 10+ times, Crime 15+ times, Hell 7+ times, and Border 15+ times.
This is not a single binary. It is a menu of independent Yes/No questions bound to one event.
How the resolution rule constrains everything
Read the small print and one line does most of the work. The market resolves only on what Trump says during the scheduled New York remarks on 14 August 2026. A speech he gives the day before, a Truth Social post that morning, a rally in another state that week, a press gaggle on the way out: none of it counts. The contract is a sealed box around one appearance.
That makes the market unusually vulnerable to two things. The first is the format of the remarks. A short set of prepared statements produces very different word frequencies from an hour of improvisation, and the count-based rungs (a listed term "7+ times" or "15+ times") depend heavily on which of those the day turns out to be. The second is the event itself. If the appearance is definitively cancelled or does not qualify under the rule, the "-No Qualifying Event-" outcome is the one that pays. Everything else, on that scenario, resolves No.
That is what a resolution rule of this shape does. It converts a political day into a checklist, and the checklist is the whole game. If you have not worked through what market resolution actually decides in a prediction market, a contract like this one is a useful place to look, because every outcome hangs on a single reviewer applying a single rule to a single transcript.
Reading the ladder without over-reading it
Polymarket displays each listed term at its own implied probability. Every price below is from the same snapshot, taken on 13 August 2026, and each is a snapshot rather than a settled fact.
The single-mention rungs sort roughly as you would expect from a Trump speech-word market. On 13 August, Biden sat at the top of the board at 87.5%, Communist / Communism at 80.5%, Oil / Gas / Gasoline at 79.5%, Police Officer at 79%, Hottest at 75.5%, Stock Market at 69.5% and Crazy at 67.5%. Further down on the same date, Transgender was at 56%, MAGA / Make America Great Again at 55.5%, Through The Roof at 45.5%, Supreme Court and Zohran / Mamdani both at 39.5%, Kamala at 36.5%, Barack Hussein Obama at 28% and "Think About It" at 27.5%. Bitcoin / Crypto / Cryptocurrency was the tail of the board at 2.35% on 13 August, with "-No Qualifying Event-" just above it at 3.25%.
The count rungs are the more interesting half of the board, and on 13 August they priced like this: Drug / Murderer 10+ times at 56.5%, Crime 15+ times at 43%, Hell 7+ times at 39.5%, Border 15+ times at 36%.
Put one of those beside the single-mention rungs and the whole point of a word market falls out of it. Drug / Murderer 10+ times, at 56.5% on 13 August, prices above Through The Roof at 45.5%, above Kamala at 36.5%, above Barack Hussein Obama at 28%. The market thinks he is more likely to say one word ten times than to say another word once. That is not a claim about what the speech is about. It is a claim about how this particular speaker uses language: certain words arrive in clusters when a theme is running, while others are single set-piece references that either come up or do not. A thematic reading of the board cannot produce that ordering. A literal one can.
The temptation, looking at this, is to read the ladder as a portrait of what Trump is likely to talk about. That is fine as a party game and useful as a rough guide, but the contract does not ask what he will talk about. It asks whether specific strings of words appear. A speech that is entirely about crime, in tone, can still miss the "Crime 15+ times" rung if he circles the theme without hitting the count. A speech that mentions Biden once, in passing, resolves the Biden rung Yes, and that rung was priced at 87.5% on 13 August regardless of whether he was the topic. Word markets reward literal reading, not thematic reading, and that gap is where the mispricing lives when it lives anywhere. Readers who want the mechanics behind how any of these prices translate to a stated chance can start with our explainer on how implied probability works on a prediction market.
What a market like this is good for
A word-count contract is not a political forecast. It is a bet on speech patterns, and its accuracy as a signal is bounded by how tightly a room of traders can model one person's verbal habits over roughly one appearance. That is a genuinely interesting thing to price, and it is the kind of market where the crowd's edge over any individual guess is real but narrow. It rewards traders who have watched a lot of Trump rallies, know his rhetorical tics, and have a view on the format of this specific event; it punishes anyone who confuses vibes about the news cycle with the literal contents of a transcript.
At iPredicta we track markets like this one for exactly that reason: they are a clean stress-test of how well a contract's resolution rule matches the question a casual reader thinks it is asking, and the gap between those two is where a lot of trading edges, and a lot of confused headlines, actually come from.
Frequently asked questions
How does the Trump New York word market decide what counts as a Yes?
The market resolves Yes on a listed term if Trump says that term during the scheduled New York remarks on 14 August 2026, and No otherwise. Count-based rungs like "Crime 15+ times" need the term used at least that many times in the same appearance. Anything said outside that event, on other days or in other formats, does not count.
What happens to the market if the New York appearance is cancelled?
There is a dedicated "-No Qualifying Event-" outcome for that scenario. If the scheduled remarks are definitively cancelled or otherwise do not qualify under the rule, that outcome is the one that resolves Yes, and every other listed term resolves No. It was trading at 3.25% on 13 August 2026, which is a snapshot rather than a live figure.
