Mantic is creating a new kind of tournament to probe the frontier of AI forecasting. Participants will compete to write forecasting questions that the best AI forecasting systems disagree on; the greater the disagreement, the greater the cash prize. There is a $25,000 USD prize pool for question-writers.

We are looking for question-writers and AI forecasters to take part.

• If you are interested in being a question-writer apply here.

• If you are interested in entering an AI forecaster apply here.

Applications are now closed.

In detail

Motivation

AI forecasting systems have improved significantly over the last few years, in part thanks to tournaments like the Metaculus FutureEval series. We think the most valuable resource that these tournaments provide is information about where further progress can be made, and this comes from seeing where the leading systems disagree. Because the forecasts include explanations, we can reason about which system was right in a particular disagreement.

At Mantic we think a lot about the problem of detecting whether a particular forecasting question was ‘hard’ or ‘impossible’. In both cases, your system scores badly. The distinction is whether there was anything you could have reasonably done to improve your forecast. In the impossible case, there wasn’t; in the hard case, there was. Detecting the difference in general is likely equivalent to solving forecasting. Detecting the difference for a specific case only requires that a different forecasting system disagrees with you and can explain why.

What we’re talking about here is the difference between aleatoric and epistemic uncertainty. Aleatoric uncertainty is the irreducible uncertainty that comes from randomness in a system, epistemic uncertainty comes from your lack of understanding of the system. No matter how well you study a fair coin, your uncertainty about the result of a flip stays the same, and that is the aleatoric uncertainty.

Where do the best forecasters disagree? The Goldilocks zone.

Until recently, disagreement between AI systems was dominated by simple errors in understanding or reasoning. However, as these systems have approached human-level forecasting, their disagreements have become more interesting: they reveal divergent models of the world or differences in their ability to gather relevant information, i.e. the differences in their epistemic uncertainty.

Unfortunately, most tournament questions do not probe these differences. Forecasters agree on both solved questions, where all that remains is the irreducible aleatoric uncertainty, and questions where no system has been able to make relative progress in reducing their epistemic uncertainty. Between these is the Goldilocks zone where there is some variation in the degree of understanding.

We think targeting the Goldilocks zone is likely to be well-aligned with pursuing high value-of-information (VOI) questions. Far from the Goldilocks zone, the forecasters add very little information, either because they make no progress or because the problem is already solved.

Core idea

Given that the level of disagreement between top forecasters determines how useful a tournament is for making progress, we think we should try optimising for disagreement. The dual tournament sets up a feedback loop that drives progress by incentivising the question-writers to discover weaknesses in the forecasters, which the forecasters are incentivised to patch. This keeps the focus tightly tracking the frontier and makes the tournament as effective as possible.

Similar dynamics are present in interspecies competition, curriculum learning, and improving security via simulated adversarial attacks or red teaming.

Tournament structure

The tournament will run for four months from the middle of April to the middle of August and there will be around 500 questions in total with some front loading of questions. We’re looking to recruit up to 50 question-writers and 20 forecasters.

All questions and forecasts will be made publicly available. Leaderboards for both question-writing and forecasting will be publicly accessible throughout the tournament. This includes the specific questions, forecasts, resolutions, and the results of any appeals.

For forecasters

The forecasting component will follow the structure used in Metaculus AI competitions for the basics: questions will be released without warning around the clock, competitors will have an hour to submit their forecast, and spot peer scoring will be used. There will be ~no changes from the Metaculus API, so if your system is set up to participate in Metaculus tournaments then you will only need to change the base URL to forecast in this tournament.

Nonprofessional forecasters will have their costs subsidised at $2 per forecast. A prize pool of up to $5,000 will also be distributed based on the final standings.1

Questions will be resolved automatically by LLM judges and competitors will have the option to appeal to a panel of human judges if they believe there has been an error. The panel will be selected from the question-writers and announced ahead of the start of the competition.

For question-writers

Each question-writer will be allocated a budget of questions and given access to a question submission flow that includes AI assistants to help refine and improve their questions. Questions will be automatically screened for general suitability to avoid trivial or ambiguous questions. Submitted questions will be queued for forecasting by the AI systems and once the forecasting is complete, questions will be scored based on the degree of disagreement between the forecasters. This means question writers will get fast feedback on the quality of their submissions and the question-writing leaderboard will be live-updating. See our technical appendix for more detail on the scoring methodology.

The maximum disagreement score that can be achieved is a little less than 500 points and we will pay $10 for every 100 points. Any prize money that remains at the end of the tournament will be distributed in proportion to the final leaderboard standings.

Applying

We are looking for question-writers and AI forecasters to take part.

• If you are interested in being a question-writer apply here.

• If you are interested in entering an AI forecaster apply here.

At Mantic, we are building a general AI superforecaster to predict the world’s most important issues. We currently have the strongest AI track record for prediction accuracy, placing 4th in a recent major forecasting tournament against human forecasters, which was covered by The Atlantic.1

The Iran crisis is therefore a live test of our system – this is the first major geopolitical crisis where advanced AI forecasting capabilities are available to generate predictions in real time.

Mantic has been built to get predictions right, update them frequently, and automatically expand the question space as events unfold. So far, we’ve performed well across these three dimensions:

1. Accuracy: on the questions that have resolved, Mantic’s forecasts held up well against outcomes and prediction markets.

2. Speed: our “tripwires”, which monitor news sources and markets and trigger fresh forecasts when relevant developments are detected, have been running around the clock as the situation evolves.

3. Scale: the system automatically generates new downstream forecasting questions as the crisis cascades into energy, shipping, and macro domains.

We break down the details below.

How has Mantic performed so far

The US strike

Prior to the initial attack on February 28th, Mantic had the probability of a US military strike against Iran at approx. 70% by mid-2026, comparable to prediction markets. The question was automatically identified and generated by the system on January 14. Unlike prediction markets, where a separate market is required for each deadline, Mantic forecasts events as a probability distribution over future dates, generating predictions of a given event happening by any date, over time, and providing a much more granular view of when something will occur.2

Case in point below. On the eve of the strikes, around midnight UTC on Feb 28, it put the probability of the event happening at 15% over the next 24 hours. By 6:17am, just minutes before the news of the strike broke, Mantic’s tripwires triggered due to price moves in markets, updating its forecast to 48% for February 28. Its final rationale cited the large scale of US naval and air force deployments to the Gulf of Oman, with escalating executive rhetoric and a perceived window of Iranian regime vulnerability. It noted that while Gulf state airspace restrictions and 70% domestic opposition provided significant headwinds, the administration’s ‘locked and loaded’ posture suggested the threshold for kinetic action had already been lowered.

Oil market disruptions

In the hours following the attack, Mantic’s forecast for the Strait of Hormuz closing to commercial shipping jumped from around 20% to 80%.3 The probability of a military attack on a commercial vessel in the Strait rose in parallel, reaching 83% by March 4, before the Safeen Prestige, a Malta-flagged container ship, became the first commercial vessel to be struck within the Strait itself.4

As the conflict escalated, Mantic started generating new downstream questions such as the probability of QatarEnergy declaring force majeure on its LNG contracts. The question was generated on March 1. On March 2, Iranian drone strikes at the Ras Laffan LNG facility led QatarEnergy to halt production, and Mantic updated its forecast significantly upwards to around 70%. On March 4, force majeure was declared.

Importantly, these questions did not have matching prediction markets. Mantic identified them as relevant and assigned accurate probabilities before any market existed to do the same.

What’s next

As the crisis persists, here is where some of our live forecasts stand as of March 24.

Ceasefire or escalation?

Ceasefire prospects between the US and Iran by June 30 stand at around 75% by mid-year, up from roughly 60% over the weekend, following signals of a possible diplomatic off-ramp from the Trump administration on March 23. Mantic identifies the spring as the most likely window, with the median date in late April: the operational timelines for Epic Fury and Lion's Roar point toward an April conclusion of the high-intensity phase, Iran's missile capacity is roughly 90% degraded, and oil above $100 is creating pressure to prevent a global recession, particularly for the Trump administration ahead of midterm elections later this year. But the path to a ceasefire is complicated by several escalation risks that are still live.

The probability of a US or Israeli strike on Kharg Island — Iran's primary oil export terminal — has fallen to 3% by end-March following the announcement of a five-day pause on strategic strikes by the US, but remains at 25% by mid-year. Mantic views the main risk window beginning after the pause if Hormuz remains constrained or Iran escalates, though it notes that strikes on the island have so far deliberately avoided oil infrastructure: the maritime blockade has already halted Iranian exports, so destroying the facility would add environmental risk and further oil supply shocks for little additional strategic gain. US officials have also discussed seizing the island rather than destroying it, a move that would defund the regime while preserving the infrastructure for any eventual transition.

Any seizure scenario would require boots on the ground. Mantic puts the odds of a US ground deployment into Iran at 28% by June 30, down from 40-45% earlier this month as the diplomatic signals have shifted. The arrival of US Marines in the Gulf still creates a near-term window for amphibious operations — whether seizing Kharg Island or neutralizing other coastal infrastructure — but the administration’s recent rhetoric suggests a preference for declaring victory through air power and economic pressure rather than a ground campaign. A majority of US voters oppose a ground invasion and broader coalition support from key allies remains thin, making anything beyond limited coastal operations unlikely without a months long build-up.

Regime change in Iran?

The Iranian regime has entered its most unstable period in decades. Mojtaba Khamenei was confirmed as Supreme Leader on March 8, but Mantic gives him a high (61%) probability of being removed by year-end, due to the combination of possible physical injuries sustained during the Feb. 28 airstrikes, active foreign targeting, and a lack of traditional clerical legitimacy.

President Pezeshkian’s position has improved: during the succession crisis, our tool had him around 60% to leave office by mid-year, but this has since fallen to around 29% after he publicly aligned his administration with the new leadership. The regime collapse and elite purge scenarios have not materialised as quickly as feared, and Mantic now views him as more likely to remain in office — though hardline consolidation under Mojtaba still leaves a meaningful probability of early departure.

Energy supply

The center of gravity is the Strait of Hormuz, where commercial shipping has effectively ceased. Mantic forecasts a 59% probability that the Strait will reopen by June 30. This is down from ~80% two weeks ago. The economic pressure for reopening is immense, with over 20% of global oil supply choked off and Brent prices above $100, the status quo is unsustainable. Our tool sees two reopening pathways: a diplomatic off-ramp, with Oman-mediated talks now reportedly underway, or a military-enforced security corridor following the degradation of Iranian naval assets. But even after hostilities subside, a full normalization of commercial traffic will require minesweeping efforts and a restoration of war risk insurance, both of which could lag a ceasefire by weeks to months.

Physical infrastructure mostly survived the beginning of the war but came under the spotlight with facilities targeted on both sides in the past week. The probability of a major attack on Saudi or UAE oil facilities causing a disruption of at least 500,000 barrels per day rose to ~80% over the weekend but has corrected back to 45-50% following the five-day pause announcement.5 QatarEnergy’s Ras Laffan facility was hit on March 18 with reports of extensive damage to key liquefaction infrastructure. The probability that QatarEnergy will lift its force majeure on LNG shipments by mid-year is now close to zero, given the scale of damage, the ongoing Strait of Hormuz blockade, and CEO al-Kaabi’s explicit condition that production requires a total cessation of hostilities.

Global economic fallout

The energy supply shock is now propagating across the global economy. Mantic places the odds of Brent crude closing at or above $120 by mid-year at 49%, down from a high of 75% a week ago. Prices have already tested this level intraday but have pulled back following the strike pause and a record IEA strategic reserve release. The supply gap from the Strait closure remains on the order of 17 million barrels per day but the outcome now hinges largely on whether the current diplomatic window produces a de-escalation or collapses back into strikes on energy infrastructure.

For global monetary policy, the crisis presents a stagflationary dilemma. In the US, a weakening labor market would normally invite cuts but the risk of energy-driven inflationary pressures mean that the probability of a Fed rate cut at the June meeting has fallen to 10%. Prediction markets seem to agree, and Mantic has generally put a higher probability than markets on this hawkish tilt over the last weeks. The ECB’s position is starker. The probability of a rate cut in June sits at just 3%, with a >70% probability of a hike. The energy price shock has effectively ended the easing cycle in Europe.

Navigating an uncertain world

Most of the questions in this post did not exist a month ago. As the crisis cascaded from airstrikes to energy disruptions to monetary policy, Mantic generated them, assigned probabilities, and monitored for updates 24/7, in real time, in many cases where no prediction market existed. The world is becoming more complex and more fragmented. We are building to help decision makers navigate this uncertainty.

We are accepting commercial clients — reach out at mantic.com