Exponential Time Hypothesis disproof confirmed by...?
This event asks whether a disproof of the Exponential Time Hypothesis (ETH), the conjecture that 3-SAT cannot be solved in subexponential time, gets formally confirmed by a set date. OpenAI says its AI system disproved ETH and that the result was checked in Lean, but the rules require confirmation from outside OpenAI or from peer review. OpenAI's own claim, or a Lean file that merely compiles, does not count.
Leading right now: June 30 86%, December 31 54%.
| Market | Chance of yes | Deadline |
|---|---|---|
| Exponential Time Hypothesis disproof confirmed by June 30, 2027? | 85.5% | Jul 1, 2027, 3:59 UTC |
| Exponential Time Hypothesis disproof confirmed by December 31, 2026? | 54.0% | Jan 1, 2027, 4:59 UTC |
Odds read from Polymarket at 22:18 UTC on Nov 21, 2026. $44K traded so far, $235 in the last 24 hours.
How this market resolves
Yes: By the deadline of the chosen market, at least one qualifying confirmation must exist. Either (i) a paper presenting a disproof is published or accepted in a reputable peer-reviewed journal or conference such as STOC, FOCS or SODA. Or (ii) a researcher or group not affiliated with OpenAI publicly states without qualification that the disproof is correct or that ETH is false. That includes Impagliazzo, Paturi, Williams, their review group, or any other credible unaffiliated expert.
No: The market resolves No if no qualifying confirmation is made public by the deadline. This includes the case where review is still ongoing, only hedged statements exist ('no errors found so far', 'appears likely correct'), or the only support comes from OpenAI or media relaying OpenAI. A confirmation made public after the deadline does not count, even if the review was done earlier.
When it really ends
Both markets end at 11:59 PM ET on their named date. The December 31, 2026 market ends at 11:59 PM EST on Dec 31, 2026, which is 2027-01-01 04:59 UTC. The June 30, 2027 market ends at 11:59 PM EDT on Jun 30, 2027, which is 2027-07-01 03:59 UTC. The endDate fields in the data match these moments.
What counts
- A paper presenting a disproof, by OpenAI or anyone else, that is published in or accepted by a reputable peer-reviewed journal or conference
- A public, unqualified statement from Impagliazzo, Paturi or Williams (or their review group) that the disproof is correct or that ETH is false
- An unqualified confirmation from another credible researcher or group unaffiliated with OpenAI, posted on a paper, preprint, website, blog or official social account
- A confirmation based on the Lean formalization, if it states that the formalized result establishes that ETH is false
- A corrected, revised or extended version of the OpenAI proof, or a new disproof by others, as the thing being confirmed
What does not count
- Preprints and other work that has not finished peer review, as a route under method (i)
- OpenAI statements, posts or press releases, or media that only relay them
- The Lean code published by OpenAI, or a check showing only that it compiles
- Statements that review is ongoing, that no errors have been found so far, or that the proof seems likely correct
- A faster 3-SAT algorithm or other weaker result that does not refute ETH
- Confirmations made public after the deadline
Fine print people miss
- The rules text sets the deadline, 11:59 PM ET, not the endDate field, though here the two line up once converted to UTC.
- A confirmation must be published by the verifying researcher or group, not just relayed by OpenAI.
- The review group's earlier statement that its involvement is not a certification does not block a later qualifying confirmation from it or its members.
- OpenAI's employees, contractors and investors count as affiliated; the Impagliazzo-Paturi-Williams group does not.
- Resolution rests on a consensus of credible sources and experts that the confirmation is legitimate, which leaves some room for judgment.
- The rules do not say how many confirmations are needed: one qualifying method is enough.
What moves the price
- A named expert such as Williams, Impagliazzo or Paturi posting that the proof is checked and correct would push toward Yes
- Acceptance of the paper at a major theory conference or journal would push toward Yes
- A credible researcher finding a gap or error in the proof would push toward No
- Experts saying only that the Lean formalization is correct but its statement does not capture ETH would push toward No
- Continued silence or only hedged comments from outside experts as the deadline nears would push toward No, especially for the earlier market
- Slow peer review cycles make publication by the end of 2026 unlikely, while the later date leaves more room for it
Put a bot on this market
A PROWL bot reads the news around the clock, judges each post against these exact rules, and trades inside the limits you set. It starts in paper mode.
Questions people ask
What is the Exponential Time Hypothesis?
It is a conjecture by Russell Impagliazzo and Ramamohan Paturi that 3-SAT cannot be solved in subexponential time. Many results in complexity theory are built on it.
What did OpenAI claim?
On October 26, 2026, OpenAI announced that its AI system proved ETH false and that the result was verified in Lean. The market asks whether that gets confirmed in a way the rules accept.
Does OpenAI's Lean proof settle the market?
No. The Lean code published by OpenAI does not qualify on its own. An unaffiliated expert would have to state that the formalized result establishes that ETH is false.
Does a preprint count as confirmation?
Not under the peer-review route. A preprint can carry a qualifying expert confirmation under the independent-verification route, if its author states without qualification that the disproof is correct.
When do the two markets close?
The December 31, 2026 market closes at 11:59 PM ET that day, which is 2027-01-01 04:59 UTC. The June 30, 2027 market closes at 11:59 PM ET, which is 2027-07-01 03:59 UTC.
Would a statement like 'no errors found so far' be enough?
No. The rules require an unqualified statement that the disproof is correct or that ETH is false. Statements that review is ongoing or that the proof looks likely correct do not qualify.
More Science markets explained
Written from the market's own rules text on Nov 17, 2026. If the rules on Polymarket differ from this page, the rules on Polymarket govern.