The Oracle
Ask the Oracle.
Ask any resolvable question about a future event and get back a calibrated probability — not a guess, a number you can hold to account. Every answer is reasoned by an ensemble of frontier models, weighed by a supervising model, then published publicly and stamped to the Bitcoin blockchain so it can never be quietly changed after the fact.
$5 per question · every answer public · every answer timestamped
The Oracle wants a sharper question
Suggested wording
The Oracle will answer exactly this question
Once asked, an ensemble of frontier models deliberates for ~8 minutes, then the answer is published to the public record and timestamped.
Your question is held
Paid questions open soon. When they do, this one answers for $5 — an eight-model ensemble, deliberated and timestamped. Consider it reserved.
The Oracle declines
The Oracle answers questions about the future. It does not take instructions.
How to ask a good question
The Oracle answers questions that reality can settle. Three things make one resolvable:
- A clear event. One specific thing that either happens or doesn’t — no vibes, no “will AI change everything.”
- Explicit resolution criteria. Say exactly what counts as yes. Name the source that decides it if there’s any doubt.
- A date. Give the question a deadline, so the answer can be scored when the day comes.
Good: “Will SpaceX land humans on the Moon before 1 Jan 2028, per an official NASA or SpaceX announcement?”
Proof of accuracy
The Oracle has a track record you can check. These numbers come straight from the resolved record below — they move honestly as questions settle.
The public record
Every question the Oracle has answered on Polymarket markets — open bets first, then everything that has since resolved. Nothing is ever taken down.
Resolved
Will Christian Estrosi win the 2026 Nice mayoral election?
All 8 agents converge on Estrosi being the underdog despite incumbency. The critical evidence is not just the 10-point polling deficit but the 42% base defection rate — this is not a normal incumbency challenge where historical base rates apply. Agents 4 and 5 provided the strongest evidence by identifying Chesnel-Le Roux's explicit refusal to withdraw (making triangulaire likely, which...
2026 U.S. House Election: Republican Odds under 15% by March 31?
Two agents (2,5) failed research and should be excluded. Remaining 6 agents cluster 22-35% with effective mean ~30.3%. The critical finding is that Republican odds are already at ~18%, very close to the 15% threshold. However, the 4-hour sustained window requirement is a meaningful barrier, and the sibling market structure (under 10% at 3%, over 25% at 95%) strongly implies...
2026 U.S. Senate Election: Republican Odds over 75% by March 31?
Seven of eight agents converge tightly in the 3-7% range, with strong agreement on the key facts: (1) underlying Republican Senate odds are approximately 60-64% and declining from 68% a month ago, (2) a 11-15 point jump would be needed to reach 75%, (3) the political environment is trending against Republicans (Trump approval ~37-42%, ICE controversies, generic ballot D+4-6), (4)...
2026 U.S. Senate Election: Republican Odds under 55% by March 31?
Discarding Agent 1 (failed research) and Agent 2 (93% - clearly misread the underlying price level), the remaining 6 agents cluster at 23-38% with median ~24.5%. The sibling market structure is the strongest evidence: under-60% at 76% but under-55% at only 21% implies informed traders see a hard resistance zone around 55-60%, consistent with the structural Republican Senate advantage (53-47...
Fact Check: is L.A. U-Haul attack perp a U.S. Citizen?
Seven of eight agents with meaningful research converge on 10-22% probability, with strong consensus on the key dynamics. Agent 5 (72%) is a clear outlier that over-weighted the historical base rate (>90% of domestic attackers are citizens) while under-weighting the critical resolution mechanism: this market doesn't ask WHETHER he's a citizen, it asks whether citizenship will be CONFIRMED by official...
Kang Sun-woo in jail by March 31?
All 8 agents agree on the core factual situation and identify the same sequential hurdles. The spread (52-82%) comes from different weighting of Assembly vote passage probability and court warrant issuance likelihood. Agent 3's low estimate (52%) is partly driven by the irrelevant 4.7% executive imprisonment base rate (wrong reference class - this is pretrial detention, not post-conviction). Agent 1's...
Nothing Ever Happens: Elon Edition
Discarding Agent 8 (failed research, 0 confidence), the remaining 7 agents cluster tightly at 85-89% with strong agreement on the ranking of risks. All identify deportation and admin rejoining as near-zero, trillionaire as low (5-8%), and baby as the primary risk (5-10%). The community discussion confirms this view ('might as well rename this will elon musk have another baby'). My...
Pam Bondi out as Attorney General by March 31?
All 8 agents converge tightly (3-9%, mean 6.25%) on a low probability, representing strong consensus. The key evidence is overwhelming: (1) Trump publicly defended Bondi as 'fantastic' - and only Trump can fire her; (2) historical base rate for AG departure within first year, let alone 6 weeks, is effectively 0%; (3) Bondi shows no signs of wanting to resign;...
Will Argentina’s monthly inflation in February 2026 be between 2.2% and 2.4%?
Seven of eight agents (excluding Agent 1 which failed) converge strongly on 11-18% probability, with the median at 14% matching the market price. Agent 1 (50%) is clearly an artifact of a failed research process and should be discarded entirely. The remaining agents show remarkable consensus on the key evidence: (1) January 2026 was 2.9%, continuing a 5-month acceleration trend;...
Will Canada’s February 2026 unemployment rate be 6.5%?
7 of 8 agents estimate below market price, with strong consensus on key facts: January's 6.5% was artificially driven by labor force exit, not employment strength. The precision requirement (landing in 6.45-6.54% band) combined with recent 0.3pp monthly volatility makes exact repetition unlikely. Trading Economics professional forecast of 6.7% suggests upward bias. Sibling market analysis (summing to 209%) confirms systematic...
Will CDU win the second most seats in the 2026 Rhineland-Palatinate parliamentary elections?
All 8 agents agree on the core facts: CDU leads polls by 3 points and is projected first, not second. For CDU to finish second, SPD must overtake them - plausible given SPD's historical outperformance in RLP and Schweitzer's popularity, but against the consistent polling trend. The ensemble mean (32.1%) essentially matches the market price (32.0%), and agents show moderate...
Will Chuck Schumer vote "Yea" on the Department of Homeland Security Appropriations Act, 2026?
7 of 8 agents converge on 55-70% range with similar evidence and reasoning. Agent 1 (12%) is a clear outlier likely misreading the situation. The question decomposes into P(vote occurs) × P(Schumer votes Yea | vote occurs). If a bipartisan deal passes, Schumer as lead negotiator votes Yea with ~90%+ probability. The key uncertainty is whether a deal happens by...
Will Cornyn flip Paxton for Texas Rep Senate Primary Winner by March 2?
All 8 agents converge on 15-23% with identical evidence and reasoning. The core issue is that Paxton leads the underlying Polymarket by ~45 points, and the only realistic catalyst (Trump endorsement of Cornyn) appears unlikely given Trump's stated preference to back winners and reluctance to endorse. Early voting starting Feb 17 further reduces the window for dramatic shifts. The market...
Will Democrats 66 (D66) win the most seats in The Hague’s 2026 municipal election?
All 8 agents agree on the core facts: D66 trails Hart voor Den Haag by 4 seats (8 vs 12) in the most recent polling, with the gap exceeding the margin of error. All agents identify the same evidence structure - strong polling evidence against D66, with only weak-to-moderate upside factors (national momentum, undecided voters, margin of error). The range...
Will Dick Durbin vote "Yea" on the Department of Homeland Security Appropriations Act, 2026?
Six of eight agents produced substantive research and converged strongly on 22-28%, with two failed agents defaulting to 50% (pulling the mean up artificially). Excluding the two failed agents, the effective mean is ~26%, and the median of all agents is 28%. The consensus is remarkably strong: every substantive agent found the same core evidence — Durbin voted Nay on...
Will Dropbox (DBX) beat quarterly earnings?
All 8 agents agree on the core thesis: Dropbox has a near-perfect recent beat streak with substantial margins, the consensus bar is modest, and the SaaS model provides predictability. The two lower agents (82-83%) over-weight sentiment signals like insider selling and stock price weakness, which are poor predictors of binary earnings outcomes. The higher agents (90-91%) correctly emphasize the mechanical...
Will February 2026 be the 4th or lower hottest on record?
7/8 agents agree Feb 2026 will be 4th or lower, with the primary disagreement being confidence level rather than direction. Agents 1 and 7 directly accessed the resolution source and found Feb 2026 at 108, which ranks 6th. Even skeptical agents note that reaching 3rd place (126) would require an unprecedented ~0.18°C jump from current trajectory during La Niña/neutral conditions....
Will Jace Yarbrough be the Republican Nominee for TX-32?
Six substantive agents converge on 72-80% (mean 75.8%). Two agents returned 50% due to research failures and should be excluded. All substantive agents agree on the key facts: Trump+Abbott endorsements, 9-candidate field requiring runoff, Yarbrough as front-runner, Binkley as main competitor. The disagreement is narrow — how much to discount for runoff risk. Historical Trump endorsement win rates (83%+) in...
Will Jessica Steinmann be the Republican Nominee for TX-08?
All 8 agents agree Steinmann is the clear frontrunner, ranging from 73-85% probability. The evidence is remarkably consistent across agents: (1) Cook Political Report's 'hers to lose' assessment, (2) dominant endorsement portfolio from the entire Texas GOP establishment, (3) Super PAC support from Winning For Women. Agent 2 (85%) appears slightly overconfident by underweighting the 6-way field/runoff risk and Jensen's...
Will John Carter be the Republican nominee for TX-31?
All 8 agents agree Carter is the heavy favorite (range 85-96%, mean 87.6%). The core case is strong: 12-term incumbent, Trump endorsement, fragmented weak opposition (Gomez banned/registration suspended), and >95% historical base rate for incumbents. The main risk is the declining vote share trend (82→71→65%) potentially pushing him below 50% with 9 challengers, triggering a Texas runoff. However, even in...
Will NVIDIA (NVDA) beat quarterly earnings?
All 8 agents agree directionally that NVIDIA is highly likely to beat, with estimates ranging from 89% to 93% (very tight 4pp range, std dev 1.5%). This is strong consensus. The core evidence is robust and multiply-verified: (1) 90%+ historical beat rate over 20 quarters, (2) 12 consecutive quarter beat streak, (3) strong Q4 guidance of $65B ± 2% revenue...
Will Patty Murray vote "Yea" on the Department of Homeland Security Appropriations Act, 2026?
Agents unanimously found Murray's explicit opposition to the current bill and her Nay vote on cloture. They split on whether a compromise will be reached in time. The key insight is this is really two questions: will a vote happen, and will Murray support it. Her institutional role means she'd likely support any deal she negotiates, but the current impasse...
Will Raja Krishnamoorthi be the Democratic nominee for Senate in Illinois?
All 8 agents agree Krishnamoorthi is favored (range 60-73%, mean 69%). The evidence is specific and verifiable: consistent double-digit polling leads across 10+ polls, massive fundraising advantage, and early voting already underway. The market at 56% implies near-parity which contradicts all available polling data. The strongest counter-arguments (Stratton momentum, Pritzker money, endorsement gap, undecideds) are real but insufficient to justify...
Will Ron Eller be the Republican Nominee for MS-02?
Seven of eight agents cluster tightly between 28-34%, with only Agent 1 (50%, failed research) and Agent 5 (45%) as outliers. Agent 5's higher estimate is based on claimed establishment endorsements (Wicker, Hyde-Smith, Kelly, Reeves) that no other agent corroborated - these may be from the 2024 cycle rather than 2026. The strongest evidence comes from Agents 2, 3, 7,...
Will Russia capture Hryshyne by March 31, 2026?
All 8 agents converge tightly (20-28%, σ=3%) around ~22-24%. The evidence is remarkably consistent: Russia has a foothold in southern Hryshyne but the target intersection is in the center, ISW shows a multi-day stall with repelled assaults, advance rates have slowed, and Ukrainian defenses are holding. The market has already corrected sharply from 36.5% to 25% over 4 days, pricing...
Will Russia enter Khatnie by February 28, 2026?
All 8 agents converge tightly (4-9%, mean 6.1%, median 6.0%, std dev 1.6%) on a low probability. The consensus is remarkably strong: every agent found the same core evidence - ISW reports Russian forces attacking 'near Khatnie' daily but consistently failing to advance in the Velykyi Burluk direction. Key consensus findings: (1) 3+ months of failed attacks near Khatnie, (2)...
Will Ryan Gosling attend the Oscars?
7 of 8 agents agree on core facts: not nominated, no announced role, promotional incentive exists but is speculative. Agent 5 (82%) is a clear outlier with poorly sourced claims about promotional attendance patterns and should be heavily discounted. Removing Agent 5, the remaining 7 agents average ~33% with median 28%, very close to market price. The market dropped from...
Will the Bank of Mexico announce a decrease at the March meeting?
Agents split into two camps: those emphasizing hawkish signals (pause, inflation, BBVA guidance) at 28-42%, and those giving more weight to Heath's board dynamics revelation and BofA forecast at 45-50%. The ensemble mean of 41.5% vs market price of 49.5% suggests a potential edge on NO, but the edge is only ~6.5% and the confidence is very low (std dev...
Will the NYC nurses strike end by February 28, 2026?
All 8 agents independently found the same critical facts and all estimated below market price (range 28-62%, mean 45.4% vs market 75.4%). The evidence is specific, verifiable, and strongly directional: (1) 73% rejection of the last deal shows massive gap between offer and demands, (2) no talks even scheduled with 12 days left, (3) NYP is the most anti-union system...
Will the U.S. flu hospitalization rate per 100,000 in Week 6 be between 60 and 70?
Agents split into two camps: three Claude Opus agents (4,5,6) at ~40-42% based on strong increment trend analysis showing 3.7-3.8 stable increments making it likely to exceed 70; three other agents (2,3,8) at 55-62% emphasizing declining flu A. The Opus agents had the strongest specific evidence — they identified the backfill effect (weekly rate of 2.2 but cumulative increment of...
Will there be no change in Fed interest rates after the March 2026 meeting?
All 8 agents converge tightly on 89-95% probability of a hold, with a mean of 92.9% and low standard deviation (1.8%). This is a strong consensus. Every agent identified the same core evidence: (1) CME FedWatch at 90-94% hold probability, (2) Fed held in January 2026 establishing a pause pattern, (3) strong January jobs report (130K vs 66K expected), (4)...
Will Trump say "Antifa" in February?
Agents 1 and 2 had research failures and should be discounted. The remaining 6 agents range from 35-55%, splitting on how much weight to give the SOTU. The strongest evidence comes from Agent 4 and Agent 8 who verified multiple transcripts showing no February mentions. The SOTU is a real but uncertain catalyst - Trump's speeches there are scripted and...
Will Trump say “Low IQ” by February 28?
Agents split into three camps: high-confidence YES (Agents 1,4,7 at 85-99%) based on specific evidence claims that may predate market creation; moderate YES (Agents 2,5,8 at 72-75%) based on base rates; and near-coin-flip (Agents 3,6 at 48-55%) based on transcript checking finding nothing in February. The base rate argument is compelling - ~1 use/week over 12 days gives roughly 82%...