Geopolitical Forecasting [GF] Challenge

Can you create a method to predict future geopolitical outcomes that is better than current state-of-the-art methods?

This challenge is closed

Fall 2018: Winner Announcement

This challenge is closed


Challenge Overview

The Problem

Decision makers rely on the Intelligence Community to provide accurate and relevant forecasts, and IARPA is working to identify methods to maximize the quality of these forecasts. IARPA’s Geopolitical Forecasting (GF) Challenge invites individuals or teams to develop innovative solutions and methods for integrating crowdsourced forecasts and other data into accurate, timely forecasts of geopolitical events. Solvers will be competing against each other and in parallel with a similarly structured research program funded by IARPA. Over the course of approximately seven months, the GF Challenge solvers will be asked to develop solutions capable of producing forecasts to a series of questions like:

  • Who will win the upcoming presidential election in Egypt?
  • or What will the spot price of Brent Crude oil be on [date]?

In order to answer these questions, solvers will be given access to a continuously updated stream of forecast judgments produced by a crowd of human forecasters and will be allowed to use other data sources to produce their solutions. The challenge presents an opportunity for individuals and teams to earn prizes by creating methods that successfully forecast of a wide variety of geopolitical events, such as political elections, international conflict, disease outbreaks, and macro-economic indicators. 

Challenge Details

Solvers will be invited to the GF Challenge Platform to compete. Each solver or team will be assigned a unique application programming interface (API) token and login account for the platform. A steady stream of questions (roughly 25 per month) will be released for solvers to produce probabilistic forecasts against (specific requirements will be described in the rules). Questions may be binary (yes/no), multiple choice, or ordered outcome (binned quantity/date). In addition to the questions, solvers will receive access to a continuously updated stream of forecast judgments produced by a crowd of human forecasters. Solvers are encouraged to create solutions that use this stream of human judgments alongside other publicly available data streams and information to create their forecasting solutions.

  • Questions will be released on the platform and can be pulled through the API
  • Each forecast question will include a set of exhaustive, and mutually exclusive response options, and will have a final resolution date that may range from a day to many months in the future. Challenge participants will be asked to produce forecasts until the question resolves. 
  • Examples of potential questions*:
  • Will the WHO confirm more than 10 cases of Marburg in 2018? (Yes/No)
  • Before March 2018, with South Korea file a World Trade Organization dispute against the United States related to solar panels? (Yes/No)
  • All forecasts will need to be submitted through the API
  • Forecasts may be updated and will be scored on a daily basis for each question that is open.


Scores will be calculated using a metric which incorporates forecast accuracy and confidence. This metric is based on the Brier score, a quadratic measure of the forecaster's closeness to ground truth for the outcome that occurs. For example, an accurate forecast with 90% confidence for an event that occurs will score four times better than a forecast with 80% confidence. The score will be calculated on a daily basis over the duration of the question; those daily accuracy scores will then be averaged. The more accurate and timely the forecast, the better score.

As questions will range in duration, milestone check-ins for questions resolved during a certain period will be used to award progress-based incentives. This means solvers who start after the challenge has launched are still able to compete for some of the prizes. Finalists who wish to compete for monetary prizes will need to provide the solution package for review the GF Challenge Team at the end of the challenge, or as appropriate, if requested. Solutions will need to be documented and all data streams used in prize-winning solutions will need to be outlined and verified by the GF Challenge Team. Please refer to guidelines for a description of scoring.


Solvers will be able to compete on their own or form teams to compete for a total prize purse of $200,000. The prize amounts shown are dependent upon certain criteria and/or participation requirements. Challenge results and solutions will be reviewed at an IARPA forecasting Workshop in Washington, D.C. For additional details and prize requirements, please refer to the guidelines. Prizes will be distributed for the following criteria:

  • Overall Performance Prizes
    • 1st Place - $20,000
      • 1st Place could be eligible to receive the Performance Bonus: Ultimate Forecaster, an additional $40,000*
    • 2nd Place - $15,000
    • 3rd Place - $10,000
    • 4th Place - $7,000
    • 5th Place - $4,000
      • Second through Fifth Place may be eligible for a share of Performance Bonus: Star Forecasters, an additional $30,000*
  • Additional Incentives
    • Best in Domain / Region Pair (X5) - $25,000 total
    • Best Undergraduate - $4,000
    • Milestone Performance (X3) - $30,000 total
    • Interim Prizes - up to $15,000

*Bonuses have additional eligibility requirements. Please see the challenge rules for details.

Rules and Eligibility

Individuals and teams, 18 years of age or older will be able to compete for cash prizes. Some individuals or organizations may not be eligible for monetary prizes. However, these solvers may, upon IARPA approval, participate in the challenge and be eligible for ranking on the leaderboard. Approval will need to be obtained in advance of participation by emailing:


What Can You Do Right Now?

  • Click “Follow” to receive challenge updates and follow its progress!

  • Register for the challenge by clicking "Accept the Challenge" and agreeing to the challenge rules and IP agreement.
  • Share this challenge on social media using the icons above. Show your friends, your family, or anyone you know who has a passion for discovery.
  • Start a conversation in our Forum to join the conversation, ask questions or connect with other innovators.

Challenge Guidelines

Challenge Rules Document 8.31.18 (Updated)

Challenge FAQ's 2.21.18 (Updated)

Final Submission Document 8.23.18



Can you create a method to forecast the future? The Geopolitical Forecasting (GF) Challenge invites Solvers from around the world to develop innovative solutions and methods for integrating crowdsourced forecasts and other data into accurate, timely forecasts on worldwide issues. The challenge presents an opportunity for individuals and teams to earn prizes by creating methods that successfully demonstrate a forecast of a wide variety of geopolitical events, such as political elections, disease outbreaks, and macro-economic indicators.


Existing methods of geopolitical forecasting include human judgment-intensive methods, such as prediction markets, and data-intensive approaches, such as statistical models. GF Challenge Solvers will develop solutions that produce probabilistic forecasts in response to numerous closed-ended forecasting questions that concern specific, objectively verifiable geopolitical events containing timeframes with deadlines and locations. The effort will run in parallel to IARPA’s most current geopolitical forecasting research program-- Hybrid Forecasting Competition (HFC). Challenge Solvers will be competing on a largely overlapping set of Individual Forecasting Problems (IFPs) as HFC research teams, and given access to the same human forecaster data stream. In addition to the provided data stream, Solvers may use other data streams and their own data and models for the challenge (in accordance with applicable laws).

Challenge Process

The following three (3) stages are anticipated as part of the GF Challenge:

Stage 1: Registration and Configuration

Solvers will register on the HeroX platform and acknowledge their understanding and intent to follow the rules of the challenge. Following registration, instructions will be sent to access the application programming interface (API) within the Cultivate Labs platform. Solvers will be granted an API token and documentation for interacting with the challenge platform.
Stage 1 registrations will be accepted on a rolling basis, so Solvers may register at any point. However, solvers will be rewarded by accuracy scores that incentivize participation on their ability to forecast on more IFPs. Eligibility for prizes (documented below) have minimum participation requirements for the resolved IFPs included in each prize.

Stage 2: Development, Forecasting and Scoring

Forecasting questions, known as IFPs, will be released periodically and will be open for a specified period of time. Solvers should check daily for new IFPs. While an IFP is open, Solvers may update their forecasts daily until the IFP closes. Please see the Judging and Scoring section below for the full-time periods associated with the IFP forecasting period. Each Solver will be permitted to submit forecasts using up to 25 different methodological “slots” as described in the Rules section below. Once the IFP closes, the forecasts will be locked and scores for the IFP for each method will be calculated. See the Judging and Scoring section below for further information. Scores will be tracked on the leaderboard within the Cultivate Labs platform.
In addition to the forecast submission activities, Solvers are encouraged to work to improve their forecasting methods throughout the competition.
Throughout the Challenge, there will be three (3) Milestone Check-Ins at which Milestone Prizes will be awarded. The Milestone Prizes will be awarded based on all IFPs that resolve during the Milestone duration. The Milestone schedule is as follows (milestone prizes will be validated within 30 days of the milestone end):

  • Milestone 1: March 7, 2018 – April 30, 2018
  • Milestone 2: May 1, 2018 – June 30, 2018
  • Milestone 3: July 1, 2018 – September 7, 2018

Stage 3: Solution Evaluation and Prize Awards

At the end of the challenge, the grand total scores will be confirmed and determined for each of the prize categories. Solvers will need to provide a write-up of their solution to the Government for final review and evaluation. This final review will determine if the solutions are sound and meet the criteria of the challenge. Solvers will provide a short (4 page) explanation for their solution. Before final prize award, IARPA may ask for additional documentation.
Upon conclusion of the Government evaluation of the documentation provided by the Solvers from the Evaluation stage, prizes will be awarded as follows.

Overall Performance Prizes (Based on overall score)




1st Place Overall


  • Must participate in at least 70% of the IFPs.  Must have positive NBPs on more than half of the IFPs you attempt
  • Uses the single best scoring method from all of the Solver’s available methods across all IFPs

Ultimate Forecaster


  • Must achieve the 1st Place Overall Prize
  • Must beat the GF Challenge Baseline
  • Must beat the HFC Top Score

2nd Place Overall


  • Must participate in at least 70% of the IFPs.  Must have positive NBPs on more than half of the IFPs you attempt
  • Uses the single best scoring method from all of the Solver’s available methods across all IFPs

3rd Place Overall


4th Place Overall


5th Place Overall



Star Forecasters

$30,000 split amongst all eligible Solvers/teams

  • Must achieve 2nd – 5th Place Prizes
  • Must beat the GF Challenge Baseline
  • Must beat the HFC top score
  • Bonus will be divided among the qualifying winning Solvers / teams according to the overall place as follows:
    • 4 winners: 50%, 25%, 15%, 10%
    • 3 winners: 50%, 30%, 20%
    • 2 winners: 70%, 30%
    • 1 winner: 100%

Definitions of prize criteria:

  • Net Brier Points (NBPs) – See the Judging and Scoring section point 3 for a description of Net Brier Points.
  • GF Challenge Baseline: Challenge solvers will have access to a stream of de-identified individual-level human forecast data, and an aggregate stream of human forecast data. This aggregate will be constructed using a method based on the state-of-the art from on the IARPA Aggregative Contingent Estimation (ACE) Program. The aggregate measure will serve as the baseline against which NBPs will be calculated. This baseline will be available to solvers via the Cultivate Labs platform.
  • HFC Top Score: The top scoring method coming out of the parallel HFC Perform teams will be visible on the leaderboard during the challenge.

Additional Incentives/Milestone Prizes




Best in Domain / Region Pair

$25,000 (5 awards of $5,000 each)

  • Top score on each of the five (5) domain/region pairings at the end of the challenge
  • Minimum participation in 80% of the total IFPs in the domain/region pairing
  • Uses the best scoring method to achieve the domain/region pairing, regardless of the Solver’s overall method used in final scoring

Best Undergraduate


  • Best score by an undergraduate, or all-undergraduate team who did not place in the Top five (5)
  • Minimum participation in more than half of the total challenge IFPs
  • Certification requirements (transcript, certified letter) from current institution must be presented in order to claim this prize

Milestone 1

$7,500 (10 awards of $750 each)

  • Top ten Solvers with the highest score in the Milestone will each get $750
  • Minimum participation in 80% of the total IFPs resolved within the milestone period

Milestone 2

$10,000(10 awards of $1,000 each)

  • Top ten Solvers with the highest score in the Milestone will each get $1,000
  • Minimum participation in 80% of the total IFPs resolved within the milestone period

Milestone 3


Minimum participation in 80% of the total IFPs resolved within the milestone period 

Top ten Solvers with the highest score in the Milestone will each get $1,250 

Election Forecaster$10,000
  • Top Solver in election-related IFPs at the end of the challenge
  • Minimum participation in 80% of the total IFPs in the Elections topic
  • Based on Solver's best scoring method for Elections IFPs, regardless of Solver's best overall method
Spring Forecaster$2,400 (3 awards of $800 each) 
  • Top three Solvers with the highest score on IFPs that open during the month of April and resolve by June 20
  • Minimum participation in 80% of the total IFPs in the relevant interim category
  • In the event of a tie, winning Solvers will split the prize purse evenly among the top scores
Interim Prize$2,600
  • Additional incentives will be provided throughout the challenge and at the discretion of the challenge team. Prize amounts could be applied to milestone, domain/region pairings, or additional bonus incentives to the Top 5 Overall Performance Prizes. These Interim Prizes will be announced prior to the start of the Prize Calculation.
  • Minimum participation in 80% of the total IFPs in the relevant interim category


Payment Terms

Solvers will need to submit a W-9 tax form, or a W8-BEN form in order to receive payment. Solvers are responsible for payment of all taxes incurred from the acceptance of Prize funds. Solvers will be given 30 days to submit their paperwork upon notification of award.  IARPA and Booz Allen are not responsible for lost or stolen prize payments, or incorrect routing and payment information provided by the winning Solvers.


  • All Solvers will need to fill out a registration form providing information on their team and asserting eligibility to participate through the HeroX GF Challenge page.
  • All daily system forecast submissions must be made via the Cultivate Labs GF Challenge API and authenticated using the API token provided during the registration process. All submissions must match the format described in the API document available on the Cultivate Labs Platform.
  • IFPs and the human forecast stream will only be provided to Solvers via the  Cultivate Labs GF Challenge API.
  • GF Challenge forecast days begin at 14:01 U.S. Eastern Time (Standard or Daylight Savings as appropriate). The daily reporting deadline is 14:00:59 U.S. Eastern Time.

    Each day’s final/official forecast (due 14:00:59. ET) will be treated as a prospective forecast that covers the immediately following 24-hour period, ending at 14:00:59 ET the next calendar day. The final forecast for an IFP is defined to be the last forecast submitted on the last forecasting day before the IFP is resolved. Suppose IFP closes as scheduled at 23:59:59 ET on 31 December, then a forecast at 14:00:58 ET that day would count but a forecast at 14:01:01 ET would not count. If an IFP closes early, say 09:00:00 ET on 30 December then a forecast at 14:00:58 ET on 29 December would count but a forecast at 14:01:01 ET on 29 December would not count.
  • Each Solver is permitted to submit daily forecasts for up to 25 official competition forecasting methods or permutations (aka “slots”) at any given time. Solvers shall provide names of 50 characters or fewer for their forecasting methods, which must be used consistently throughout the challenge period. These names may not include profane or derogatory terms.
  • Solvers may submit more than one forecast update per IFP per day. For each forecasting method, only the last forecast for an IFP submitted before the daily submission deadline of a given forecast day will be treated as the Solver’s official forecast for each IFP for that forecast day.
  • Each Solver’s forecasting method will be scored across all IFPs for which forecasts are submitted, and an overall forecast accuracy score for that method will be determined. For each Solver, the Testing and Evaluation (T&E) Team will identify the single method with the best score across all IFPs as the Solver’s “best” method. In other words, each Solver’s best method will be a unitary, prospectively defined method selected from among the Solver’s 25 official methods: We will not “cherry-pick” per-IFP most accurate methods to formulate composite, per-Solver “best” methods.
  • Once a Solver submits their first forecast on an IFP they should provide daily forecast updates. If they do not, their previous forecast will be carried over each forecasting day until a new forecast is submitted.
  • For forecast days prior to the first submission by a Solver on an IFP, the assumed forecast is the aggregated human forecaster stream forecast for that IFP.
  • For Milestone Prizes, only the IFPs that are resolved during that Milestone’s time period will count for that Milestone’s prize. IFPs that are opened, but not resolved, during a Milestone will not count. They will be counted towards the Milestone period in which the IFP resolves.
  • In order to be eligible for Domain/Region, Election Forecaster, Spring Forecaster, interim, and Milestone awards, a Solver must submit forecasts for at least 80% of the IFPs considered for that prize. To be eligible for the Undergrad prize, Solvers must submit forecasts for at least 50% of the total number of IFPs in the challenge and provide certification of their status as a student from their current institution. In order to be eligible for one of the Top five (5) overall performance prizes, a Solver must submit forecasts for at least 70% of the total number of IFPs in the challenge.
  • Solvers are responsible for ensuring that submitted forecasts are compliant with the standards documented within the API documents on the Cultivate Labs Platform. Forecast submissions must be proper in that the probabilities assigned to the several outcomes for an IFP must sum to 1.0. Improper, malformed or non-compliant forecast submissions will be rejected. IARPA reserves the right to change its policies with reasonable notice, including disqualification of non-compliant data.
  • Solvers may use external data streams in their solutions. However, these external data sources will need to be documented. The Solver is responsible for ensuring that they have the necessary rights and licenses for the use of such external data sources for the purposes of this challenge.
  • The Human Forecaster data stream provided through the challenge is available for Solvers to use, free of charge, only for the purposes of this challenge. All other uses are prohibited without the express consent of IARPA. However, Solvers are not required to use this data stream as part of their solution.
  • Solvers are solely responsible for ensuring they have valid licenses or rights to any software libraries or data sources used in their forecasting systems. Any method that infringes on another party’s intellectual property will be disqualified from the competition.
  • Solvers may only register once for the challenge and can compete either as an individual or as part of a team; Solvers cannot participate as both an individual and as part of a team.

Judging and Scoring

  1. Your competition score is based on how accurate your forecasts are when compared with the stream of Government-furnished human forecasts. The Government-furnished forecast stream (GFFS) is a “crowd wisdom” aggregation of asynchronously submitted human forecasts for each IFP. Given an IFP we compute the consensus forecast as a function of the weighted average of the latest forecasts from a portion of the unique forecasters who most recently provided forecasts for this IFP. The weight assigned to each forecaster is computed from three factors. The first two are based on their performance across all IFPs: historical forecast accuracy for that forecaster and frequency with which that forecaster submits updates. The third factor is whether or not they have completed a brief, self-administered forecast training course. The weighted average of the recent forecasts is then extremized toward 0 or 1 using a tuning parameter. This method for computing the consensus forecast was determined to have good performance during the IARPA ACE Program and is a suitable benchmark against which to compare solver performance.
  2. We measure absolute forecast accuracy using a formula called the Brier score, which reflects the squared difference between the probabilities a solver or the GFFS assigns to each outcome and reality (see appendix section A.1). Lower Brier scores reflect greater accuracy: A perfect score is 0 (for instance, on a binary question, you assigned 100% to the correct outcome), the worst possible score is 2 (for instance, on a binary question, you assigned 100% to an incorrect outcome). For some IFPs the possible answers have an ordering such that some answers are closer to being correct than others. In this case, we use a modified version of the Brier score that awards more credit when assigning higher forecast probabilities to answers that are closer to the correct answer (see Appendix: A.2 Application of Brier Scoring to Ordered Outcomes). The appendix provides a complete description of generic and ordered Brier scoring.
  3. For a given IFP, we compare the accuracy of your forecasts to those of the GFFS using Net Brier Points (NBPs), defined as the difference between the GFFS’s Brier score and your Brier score for each day that the IFP is open. Negative NBPs means the GFFS was more accurate than you, positive NBPs means you were more accurate than the GFFS, and more NBPs is better. For days that pass before your first forecast on the IFP, your forecast is assumed to be the same as the GFFS consensus forecast, meaning you do not accrue (or lose) NBPs for a given IFP until you submit your first forecast. Once you submit an initial forecast, that forecast holds each day thereafter until/unless you update it. Your score for the IFP is then the sum of your NBPs for all days that an IFP is open.
  4. For most of the prize categories your score is the sum of your NBPs over all germane IFPs for that category. However, the bonus Ultimate Forecaster and Star Forecasters Prizes are scored slightly differently. These bonus prizes feature a comparison of your method to the best of the performers in the IARPA HFC program on the same IFPs. For the purposes of this bonus prize category we will compute your Mean Daily Brier Score (see Appendix: A.3 Mean Daily Brier Score ) for all IFPs and take the mean of those scores across those IFPs. If your mean Mean Daily Brier score is lower than the best HFC program method you will be eligible for this bonus prize category.
  5. To be eligible for overall prizes you must attempt at least 70% of all IFPs and you must have positive NBPs on more than half of the IFPs you attempt. For example, suppose there are 150 IFPs for the overall prize category. Suppose that you submit non-trivial forecasts on 105 of these IFPs. Furthermore, let us say that you have positive NBPs on at least 53 IFPs. You will meet all of the eligibility criteria for overall prizes.
  6. In order to be eligible for a prize the Solver must document the solution method in a short document (no more than four pages) using the provided format (coming soon). Please note that IARPA will be examining the substance and clarity of the technical details documented, not the quality of the writing.
  7. Solutions competing for prizes will be verified by the Solver or Team that all data sources and software libraries can be identified through the solution with proper attribution where necessary.

Intellectual Property

Solvers will retain the Intellectual Property (IP) rights to their solutions. IARPA will use information submitted through the challenge and application process to understand the technology.

  • IARPA will have the rights to utilize all data, forecasts, and method explanations provided for U.S. Government purposes.
  • Challenge participant names, titles, general technology descriptions, photographs, and abstracts for their submissions may be utilized in challenge-related media and promotional materials or for other internal Government uses. No sensitive intellectual property information will be shared in this manner, and must be clearly marked prior to submission.
  • Challengers will need to sign the IP Agreement from within the HeroX GF Challenge page prior to participating in the challenge.


Some individuals or organizations may not be eligible for prizes for reasons listed below. However, these Solvers may, upon IARPA approval, participate in the challenge and be eligible for ranking on the leaderboard. Approval will need to be obtained in advance of participation by emailing .

Any teams who forego prizes will be publicly listed on the HeroX GF Challenge page to ensure full transparency of their competition status.

General Eligibility Requirements

To be eligible to win a prize under this competition, an individual or entity:

  1. Must have completed and submitted a registration form at HeroX GF Challenge page
  2. Must be (1) an individual or team each of whom are 18 years of age and over, or (2) an entity incorporated; and
  3. May not be a federal entity or federal employee acting within the scope of their employment. An individual or entity shall not be deemed ineligible because the individual or entity used federal facilities or consulted with federal employees during a competition if the facilities and employees are made available to all individuals and entities participating in the competition on an equitable basis.

Federal grantees may not use federal funds to develop challenge applications unless consistent with the purpose of their grant award. Federal contractors may not use federal funds from a contract to develop challenge applications or to fund efforts in support of a challenge submission.

Employees of IARPA, Booz Allen Hamilton, Cultivate Labs, HeroX, Good Judgment, Inc., and MITRE are not able to compete in the challenge. Companies and contractors who are currently supporting, or who have previously supported, the above (or other) entities in efforts related to the IARPA HFC program are also not able to participate in the challenge. This restriction extends to members of such persons’ immediate families (spouses, children, siblings, parents), and persons living in the same household as such persons, whether or not related.

Federally Funded Research & Development Centers (FFRDCs) and (DoD) University Affiliated Research Centers (UARCs) not mentioned above may be eligible to submit forecasts and receive leaderboard recognition, but are not eligible to win prize dollar awards. In order to compete, an email must be sent to with the team name, organization represented, and team member information.

IARPA’s Hybrid Forecasting Competition (HFC) performer team members and HFC Good Judgment Forecasters, as well as members of such persons’ immediate families (spouses, children, siblings, parents), and persons living in the same household as such persons, whether or not related, are not eligible to participate in the competition.

Federal employees, employees of FFRDCs, and employees of UARCs can participate in the challenge as a private citizen, not affiliated with their organization, as long as (a) the Solver is not employed by an organization that is actively involved in the challenge or closely related programs; (b) the Solver has received approval from their organization, if applicable, to participate in their personal capacity (note, it is the Solver’s responsibility to verify this); and (c) Solvers cannot use government or their organizations’ resources, computers, or access to government information in aid of their participation, except for those resources available to all other participants on an equal basis.

Entrants must agree to assume any and all risks and waive claims against the Federal Government and its related entities, except in the case of willful misconduct, for any injury, death, damage, or loss of property, revenue, or profits, whether direct, indirect, or consequential, arising from their participation in a competition, whether the injury, death, damage, or loss arises through negligence or otherwise.

Entrants must also agree to indemnify the Federal Government against third-party claims for damages arising from or related to competition activities. Entrants are not required to obtain liability insurance or demonstrate financial responsibility in order to participate in the competition.

By participating in the competition, each entrant agrees to comply with and abide by these rules and the decisions of IARPA and/or the individual judges, which shall be final and binding in all respects.

By participating in the competition, each entrant agrees to follow all applicable local, state, federal and country of residence laws and regulations.


Companies, universities, individuals, and Teams of Solvers are able to participate in this challenge. Companies / Teams will need to elect a Team Captain, who will be the main point of contact (POC) for communications. Individuals who have registered on the HeroX GF Challenge page as an individual competitor, will not be eligible to participate in a team. All Team Members must meet the General Eligibility Requirements, self-register, and acknowledge these rules through the HeroX GF Challenge page.

  • Upon registering for the challenge, the Team Captain will need to provide a breakdown of all Team Members along with the percentage allocation between all Team Members
  • Each Team Member will need to sign an agreement that they agree to distribution of prize funds
  • A Team Captain may provide information for a company or organization instead of their personal information for tax purposes.
  • One representative from the Team, will need to set up their Team’s API. There will only be one API allowed per Team for the entire duration of the challenge
  • Once a Team has submitted their first forecast, the team members cannot change. This means that no new Solvers can be added to a Team and no Solvers can leave the Team.
  • Individual Solvers who have submitted a forecast on their own may not form a Team after their first forecast has been submitted.

Foreign Nationals and International Solvers

Except for residents of Iran, Cuba, North Korea, Crimea Region of Ukraine, Sudan or Syria or other countries prohibited on the U.S. State Department’s State Sponsors of Terrorism list, Solvers residing in any country are able to participate. In addition, Solvers are not eligible to participate if they are on the Specially Designated National list promulgated and amended, from time to time, by the United States Department of the Treasury. It is the responsibility of the Solver to ensure that they are allowed to participate in this challenge and allowed to export their technology solution to the United States.

General Liability Release

By participating in the competition, each entrant hereby agrees that:

  1. IARPA and the U.S. Government, Booz Allen Hamilton, Cultivate Labs, HeroX, and MITRE shall not be responsible or liable for any losses, damages, or injuries of any kind (including death) resulting from participation in the competition or any competition-related activity, or from entrants’ acceptance, receipt, possession, use, or misuse of any prize; and
  2. Entrants will indemnify, defend, and hold harmless IARPA and the U.S. Government, Booz Allen Hamilton, Cultivate Labs, HeroX, and MITRE from and against all third-party claims, actions, or proceedings of any kind and from any and all damages, liabilities, costs, and expenses relating to or arising from entrant’s participation in the competition.

Without limiting the generality of the foregoing, IARPA and the U.S. Government, Booz Allen Hamilton, Cultivate Labs, HeroX, and MITRE are not responsible for incomplete, misdirected, prize notifications; or for late, lost, inaccessible, network connections, or for issues with Internet Service Providers, websites, or other connections and third-party sites that the entrants choose to use for the competition; or for any technical malfunctions, failures, difficulties, or other errors of any kind or nature.

These rules cannot be modified except by IARPA. All decisions by IARPA regarding adherence to these rules are final. The invalidity or unenforceability of any provision of these rules shall not affect the validity or enforceability of any other provision. In the event that any provision is determined to be invalid or otherwise unenforceable or illegal, these rules shall otherwise remain in effect and shall be construed in accordance with their terms as if the invalid or illegal provision were not contained herein.


By participating in the competition, each Solver represents, warrants, and covenants as follows:

  1. The Solver – whether an individual, team or entity – is the sole author, creator, and owner of the submission;
  2. The submission is not the subject of any actual or threatened litigation or claim;
  3. The submission does not and will not violate or infringe upon the intellectual property rights, privacy rights, publicity rights, or other legal rights of any third party;
  4. The entrant will not interfere with or attempt to influence, the events that are the subject of the IFPs in the challenge. If a Solver has a question about an activity related to their job role, please email for clarification.
  5. The Solver will not engage in activities that could cause political unrest, or lead to a violent event or civil unrest that could cause the outcome of the IFP to change.; and
  6. The Submission, and Solvers’ use of the Submission, does not and will not violate any applicable laws or regulations, including, without limitation, applicable export control laws and regulations of the U.S. and other jurisdictions.

If the Submission includes any third party works (such as third-party content, data sets, algorithms, or open source code), entrant must be able to provide, upon the request of IARPA, documentation of all appropriate licenses and releases for such third-party works. If entrant cannot provide documentation of all required licenses and releases, IARPA reserves the right to disqualify the applicable Submission, or seek to secure the licenses and releases for the benefit of IARPA, and allow the applicable Submission to remain in the Competition. IARPA also reserves all rights with respect to claims based on any damages caused by participant’s failure to obtain such licenses and releases.

Solvers – whether an individual, a team or an entity – will indemnify, defend, and hold IARPA, Booz Allen Hamilton, Cultivate Labs, HeroX and MITRE from and against all third-party claims, actions, or proceedings of any kind and from any and all damages, liabilities, costs, and expenses relating to or arising from entrant’s submission or any breach or alleged breach of any of the representations, warranties, and covenants of entrant hereunder.

IARPA reserves the right to disqualify any submission that IARPA, in its discretion, deems to violate these Rules. IARPA reserves the right to disqualify any Solver that is operating in a manner that disrupts the competition, does not comply with the spirit of the competition, or is found to be manipulating the competition. IARPA also reserves the right to amend these rules throughout the duration of the contest should extenuating circumstances arise.

Appendix (available in the downloadable Challenge Rules Document)


Challenge Updates

IARPA announces Geopolitical Forecasting Challenge 2 (GF Challenge 2) – Are you ready?

April 4, 2019, 11:01 a.m. PDT by Despina Maliaka

Are you ready?

Whether you were an active Solver or followed the first GF Challenge, we wanted to share this exciting news... GF Challenge 2 starts in May 2019!  

Check out the GF Challenge 2 page to learn more and begin engaging with this challenge.


-The GF Challenge Team

Announcing the GF Challenge Winners

Oct. 1, 2018, 5:50 p.m. PDT by Kyla Jeffrey

Over the past several months, you have participated in a unique challenge hosted by the Intelligence Advanced Research Projects Activity (IARPA). Together, we’ve watched multiple world events unfold before us and celebrated your work along the way with numerous Milestone awards. We thank you for your efforts.  


After completing the final calculations, we are thrilled to announce the overall winners of the Geopolitical Forecasting Challenge. 

1st Place: DigitalDelphi 

2nd Place: BEEFERS 

3rd Place: Catskills Research Company 

4th Place: truthLover 

5th Place: SISLers 

Ultimate Forecaster: no winner 

Star Forecaster: BEEFERS 

And here are the Additional Prize Winners in specific categories.  Congratulations! 

Election Forecaster: DigitalDelphi 

Domain Region Prizes 

  • Politics/International Relations & Middle East: truthLover 
  • Geopolitical Grab Bag & Asia: Catskills Research Company 
  • Macroeconomics/Finance & Europe: LIA 
  • Politics/International Relations & Europe: BEEFERS 
  • Politics/International Relations & Asia: truthLover

Announcing the Winners of the Workshop Presenter Prize

Sept. 25, 2018, 7:20 a.m. PDT by Kyla Jeffrey

Dear Solvers, 

It’s our pleasure to announce the winners of the GF Challenge Final Workshop.  Per the rules, these teams were selected based on Leaderboard Standings as of September 23rd.  These teams and individuals have received a presentation slot at the Final Workshop.  They will receive a $2,500 prize payment upon completion of their presentation.  

  1. DigitalDelphi
  3. catskills
  4. LIA
  5. SISLers

Congratulations – we can’t wait to meet you in the Washington, D.C. area!  

Solvers, as a reminder: While the winners of the final Workshop will receive a prize award to help with travel and accommodations, we invite all GF Solvers to join us at this special event in celebration of your participation in this unique challenge. It’s a great opportunity to network and learn from your fellow Solvers. Please keep an eye out in your email for an invitation to the workshop.  


GF Challenge Team

Milestone 3 Awards

Sept. 12, 2018, 6 a.m. PDT by Kyla Jeffrey

Dear Solvers,

It’s time to announce the winners of our third—and final—milestone! This award prize totals $12,500 and is split among the top 10 teams with 80% participation. Congratulations to the following teams and Solvers!

  2. Catskills
  3. CitizensBand
  4. DigitalDelphi
  5. LIA
  6. SISlers
  7. Seb
  8. dobreovidiu
  9. pb2pv
  10. truthLover

We’ve got more award announcements on the way, so stay tuned! You can expect to hear the final results around the end of September. Thank you so much for your participation!Cheers!

GF Challenge Team

Announcing Second Chance Awards

Aug. 31, 2018, 8:42 a.m. PDT by Liz Treadwell

As the GF Challenge enters its final weeks, we are watching the leaderboard with excitement. The Solver community has been doing an incredible job producing forecasts against a very broad portfolio of individual forecasting problems (IFPs). We are pleased to announce a new tier of “Second Chance” Prizes that may be awarded in the event that fewer than 5 methods qualify for Overall Prizes. As described in the GF Challenge Rules Document, the 1st through 5th Place Overall Prizes will be awarded for those methods with the highest Net Brier Points (NBPs), conditional on participating in at least 70% of IFPs and receiving positive NBPs on more than half of attempted IFPs. (Only each Solver’s best scoring method will be considered).

If, however, fewer than 5 winners receive Overall Prizes, IARPA will award “Second Chance” Prizes to recognize methods that achieved high NBPs, while participating in at least 70% of IFPs, but did not meet the 50% positive NBP minimum. The Second Chance award tier will be awarded as follows:

  • Methods that qualify for 1st through 5th Place Overall Prizes (and Ultimate Forecaster and Star Forecaster) will be awarded according to the originally stated criteria.
  • If fewer than 5 winning Solver-methods receive Overall Prizes, the remaining slots will become available in the Second Chance tier (e.g., if only 1st and 2nd place Overall Prizes are awarded, then 3rd through 5th place slots will be eligible for Second Chance award).
  • Second Chance prizes will be awarded using the same standards as Overall Prizes (i.e., highest NBPs, at least 70% participation) with the exception that methods will need to have positive NBPs on a minimum of only 33.3333…% of attempted IFPs.
  • Second Chance prizes will be worth half as much as their equivalent Overall Prize amounts, as shown in the table below, and will not be eligible for Ultimate Forecaster or Star Forecaster awards.
  • Power Forecaster(s) Award: Only awarded if both the Ultimate and Star Forecaster Overall Prizes are not awarded. Must participate in at least 70% of the IFPs.  Must have positive NBPs on a minimum of one third (33.3333…%) of the IFPs you attempt.  Must have a lower Brier score than the Benchmark on the Beat HFC Leaderboard.  Have a lower Brier score than all HFC Performers on the Beat HFC Leaderboard.  Uses the best scoring method, regardless of the Solver’s overall method used in final scoring. This will be awarded at $25,000 and is split amongst all eligible Solvers / teams.
  • Each Solver will only be eligible for one prize (either Overall or Second Chance) based on their single best performing method.
  • Please see the updated Rules Document for full eligibility requirements and prize breakdowns


Overall Prize

Second Chance Prizes (if applicable)




Ultimate Forecaster















Star Forecaster(s)



Power Forecaster(s)


$25,000 (split amongst all eligible Solvers / teams)


How were the percentage and prize allocations determined?

The team examined the challenge from many different perspectives—trends across time, attempt rates, success rates, as well as comparison evaluation with HFC. We recognize that we set an extremely high bar and are appreciative of the community’s achievements to date. An evaluation of the leaderboard and IFPs assisted in setting the “Second Chance” Prize percentage, which then informed the “Second Chance” prize amount awards.


Frequently Asked Questions

I don’t have an approach that I wish to enter in the GF Challenge. Is there another way that I can participate?

IARPA is interested in any method that advances the science of forecasting. This may include investigating how a combination of computer models and human judgments can be used to make accurate forecasts, but it need not. If indeed a very good human forecaster (or a small group thereof), can do just as well as hybrid solutions or the ACE baseline, then we have learned something. However, we believe this is extremely unlikely due to the number of IFPs in the Challenge; and the fact that methods will be compared, for many of the prizes, against a benchmark crowd of human forecasters using the validated state-of-the-art method from the ACE program. The administration of the Challenge is automated and secured by use of an API for IFP and baseline data distribution and forecast submission, so some programming/data science skills are necessary. The GitHub repository is our way of sharing documentation, some data, and example API code with the Solver community.

All teams must complete a prize distribution form prior to their first submission in accordance with the challenge rules. Here is the link to the distribution form. Once completed and signed by all team members, you will need to email the completed form to: .

Yes, this is a government funded project. The Intelligence Advanced Research Projects Activity (IARPA) envisions and leads high-risk, high-payoff research that delivers innovative technology for future overwhelming intelligence advantage.

In order to be fully registered for the Challenge, you will need to log into the HeroX GF Challenge page and “Accept the Challenge”. This includes accepting the Challenge rules and acknowledge that you adhere to the eligibility requirements for the Challenge, in addition to accepting the Intellectual Property (IP) Agreement. Once you have completed that step, you will be sent a link to the Cultivate Labs platform to register for your, or your team’s, API token.

HFC is an Intelligence Advanced Research Projects Activity (IARPA) program which is funding research teams to develop what we call “hybrid forecasting systems” to integrate human forecasters and statistical models in novel ways. The program is trying to harness the best of both worlds to produce forecasts that are better than either method could produce on their own. You can read more about the HFC program here:

The GF Challenge is looking for forecasting methods; solutions that are capable of processing data and making forecasts. While this may include investigating how a combination of both computer models and human judgments can be used to make accurate forecasts, it need not do so. And while your solution need not be fully algorithmic in nature, you will need to submit your forecasts throughout the Challenge through an API, and be able to describe the solution for your forecasts at the end of the Challenge, so some data science literacy is likely required.

On the other hand, if you are instead more interested in making geopolitical forecasts using your intuition, you may find participating as a forecaster in HFC ( to be more enjoyable. HFC is currently recruiting forecasters at the aforementioned URL.

Please note that you cannot participate in both the GF Challenge and HFC.

HFC teams are developing systems at a significant expense to IARPA. These teams are looking for honest experimental tests of their methodologies. IARPA is simultaneously looking for honest original solutions in the GF Challenge and we don’t want to see solutions that we are seeing in HFC. If we allowed dual-registration, there would be a possibility of Solvers accessing HFC performers’ intellectual property and using that in the Challenge. You will not be eligible for any prizes if you are registered for HFC. In order to remove yourself from participating in HFC, you need to send an email to stating your intentions and then register for the GF Challenge. By registering for the GF Challenge, you will be acknowledging that you are not concurrently participating in HFC; those who are awarded prizes will be asked to certify that they have not participated as a forecaster in HFC. 

This is a prize Challenge. This is not a traditional IARPA program. The Solvers are the ones doing the development. You are competing for prize money, not for contracting opportunities. 

For the GF Challenge, we are interested in seeing the solutions that individuals or teams can produce that are capable of intaking data and generating forecasts. To participate, you should use your product or tool to submit forecasts in response to the Individual Forecasting Problems (IFPs) that are issued during the Challenge. If you’d like more information on doing business with IARPA outside of this Challenge, you can visit the following link:

We anticipate asking between 125 to 175 IFPs over the course of the Challenge. All IFPs will be designed to resolve by the end of the competition.

You do not need to forecast against every IFP, but in order to be prize eligible, you will need to meet certain minimum targets. You can read more about the minimum participation requirements for each prize category in the prize section within the Challenge Rules Document, as well as finding information on how your responses, or non-responses, may impact your score in the Scoring section of the Challenge Rules Document.

Each prize category will have some minimum participation percentage requirements of IFPs. For the Overall Performance Prizes, it is 70% participation of all IFPs, for the milestone prizes it is 80% of the IFPs within that milestone period. Of the IFPs launched within that period of time, you must meet the minimum participation percentage associated with that prize category. When it comes to which IFPs you provide forecasts for, that is up to you. If you are thinking of using just human judgments, we discourage you from doing that. In order to submit forecasts, you must have the ability to interact with the API. If you would like to use your judgments or purely human judgments, then we recommend that you register for participation as a forecaster in IARPA’s HFC program at

Yes, every day is a scored day. The Challenge Rules Document describes what forecast days are.

Each day, there is a reporting deadline: 14:00:59 U.S Eastern Time. You will be scored by what you have as a forecast in the system at that time. We will impute or assume your forecast is the control group forecast. Once you initially submit, it will take the place of the control group forecast. 

There will be some variation between IFPs. We might ask a question on the possible outbreak of some conflict that will only be a day long. There may be some other questions where the maximum possible duration could be up to six months. The time period that a forecast will be open is a wide variation of time dependent upon the type of IFP.

Your score will be tracked on a leaderboard on the Cultivate Labs Platform. You can log into your account or Team account on the platform to see how well you are doing in the overall Challenge, or on the individual prizes/ milestones throughout the Challenge.

Once it has been determined that you have won a prize, you will be contacted via email with instructions on where to send your tax and payment information in order to receive your funds. As a reminder, all taxes that may be incurred based on the award of prize money is the sole responsibility of the Solver.

You can still participate in the Challenge! As long as you meet the minimum participation requirements for the prizes, you will still eligible to win at least some of the prizes. Note that some of the prizes have lower IFP participation requirements, and others are based on success across discrete periods of time. All is not lost if you join later in the Challenge!

You will retain rights to whatever methods you use to generate your forecasts. IARPA is supportive of further scientific research per the Office of Science and Technology Policy (OSTP) memorandum of 2013 and we would ask that you send us a courtesy copy of the publication if you cite the Challenge or IARPA within it, but you retain your IP rights, including the solutions you developed or used to make your forecasts.

The Government does not own the methods that you develop for this Challenge. We retain a copy of the forecasts that you provide and a description of your method but do not own the IP. You should have no expectation that the Government will use or others will use your methods. It is possible if you create something potentially valuable, that you may be approached from people who are interested in teaming with you on various types of efforts.

Yes, with a small number of exceptions, international Solvers are allowed to participate in this Challenge. Within the Challenge Rules Document, there is a section outlining the guidelines as to which countries are allowed to participate, and additional guidance. If you still need clarification on your eligibility to participate, please send an email to: .

No, you will need to select one of the events to participate in. IARPA’s HFC Performer team members and HFC Good Judgment Forecasters, as well as members of such persons’ immediate families (spouses, children, siblings, parents), and persons living in the same household as such persons, whether or not related, are not eligible to participate in the competition.

Yes, with a few stipulations. First, they must forego the prize money for the Challenge. Second, the entity, along with anyone working on the Challenge, must meet the other eligibility requirements in the Challenge Rules Document. Third, they may not have supported IARPA’s HFC Program in any manner. Please see the Challenge Rules Document for complete information on how to proceed.

Once the Challenge launches on HeroX, you will have an opportunity to register as a team and invite members to join the competition and join your team. However, if you decide to create a team after you have already registered, you have the ability to create a team within the community tab of the HeroX GF Challenge page. All team members must meet the general eligibility requirements, register and acknowledge the Challenge rules through the HeroX GF Challenge page. Upon registering for the Challenge, the team captain will need to provide a breakdown of all team members along with the agreed upon prize percentage allocation between all team members. Each team will need to determine who will be setting up their team’s API token, as there can only be one API token granted per Solver/Team. You are only permitted to participate in the Challenge as an individual or as part of one team (membership on multiple teams is not allowed).

No, participants can only join a single team in the Challenge and can either be on a team or be a lone Solver, but not both.

No, once you have registered for a team and your team has begun making forecasts, you cannot switch teams, or join a new team. Similarly, if you are a lone Solver and have begun making forecasts, you can no longer join a team.

Yes, individuals can form teams after the first IFP is issued, but you cannot form a team after you have submitted a forecast as an individual or the intended team has submitted a forecast. The IFPs will come out on March 7 — do not start forecasting until you have your team solidified. If you are thinking about teaming with someone, please wait to submit your forecast until you determine how you would like to participate in this Challenge (as part of a team or as an individual.)

We suggest that you post to the HeroX forum in order to seek out other team members.

Teaming presents an opportunity to divide and conquer the labor. However, many individuals prefer to wotk independently. It is up to your personal preference and working style.

These IFPs are developed by an experienced team working for IARPA who developed similar questions for the earlier IARPA Aggregative Contingent Estimation (ACE) program, which spawned Phil Tetlock’s book Superforecasting. Each IFP includes an authoritative source and/or procedure for determining the outcome. In our experience, we have found these IFPs to be well posed and to have unambiguous outcomes.

The IFPs have already been segmented into domains and regions by the IARPA team. The IFPs affect both the GF Challenge and HFC so there is not an opportunity for requests/nominations from Solvers.

Challenge Solvers will have access to a stream of de-identified individual-level human forecast data. From this, an aggregate will be constructed using a method based on the state-of-the art from IARPA’s ACE program—Solvers will also be given access to this aggregate. The aggregate measure will serve as the baseline against which NBPs will be calculated. This baseline (and the associated de-identified individual forecast data) will be available to Solvers via the Cultivate Labs API.

Each Solver is given an authentication code in the form of an API token upon registration. Forecast submissions are authenticated with this code and are not visible to others. If you keep your API token secure, you should have nothing to worry about.

All submissions must be made through the API provided by the Cultivate Labs Platform. Those wishing to make smaller numbers of forecasts using human judgment are encouraged to register for participation as a forecaster in IARPA’s HFC program at

Once you accept the Challenge Rules and IP Agreement for the Challenge on the HeroX GF Challenge page, you will be sent a link to register for your API token and access to the Cultivate Labs Platform. Please note that only one API account will be allowed for each lone Solver or team of Solvers.

No. Each team is only allowed to register for a single API account for the Challenge.

Log into your API account on the Cultivate Labs Platform and from the Home Page click the “View API Token Management” button to retrieve or re-generate an API token.

The complete API documentation will be available on the Cultivate Labs Platform upon registering for the Challenge on HeroX.

The Challenge will only provide all solvers with a single data set, which is the human forecaster data. We are not providing any additional data beyond that, as we believe this data is state-of-the-art for these types of questions. We understand that not all Solvers may want to use this data. Solvers are able to use additional or alternative data sources that they deem useful to their solutions.

The data from the ACE program is available for download in DataVerse. The ACE data - the benchmark method as well as the data that we are providing you is based on systems from IARPA’s ACE program - there will be a leader board in the Cultivate Labs site where this Challenge is being hosted. 

The final score will be based solely on the accuracy of your forecasts — you will have to turn over a document that describes your method. While the Government will gain rights to the forecasts, you will retain the rights to your IP. Your score is based on the quality of your forecasts.

We will make the exact requirements clear, but the expectation is that the document that describes your solution will be no more than four pages long. We may request more information to gain a better understanding of your method or to provide clarity. Any parts of your solution that are proprietary need to be clearly marked as such. In the event that you are a prize winner, we need to be able to describe your methodology in greater detail to a public audience so that the public can have confidence that the prize funds were awarded properly.

The results of this Challenge will be publicized through public communication channels and updated on the microsite.

Yes, but it’s quick and easy. Just click the “Join Us” button at the top of the page and follow the instructions to complete your registration. All you need to provide is your name and email address.

If you have a question not answered in the FAQ, we recommend that you post it in the Forum where someone will respond to you. This way, others who may have the same question will be able to see it.

The challenge is expected to officially launch in February 21, 2018. Please follow this challenge to get updates for when the challenge launches and more information will become available.

Once the challenge has officially launched, you will be directed to the challenge platform where you can download the API key.

For questions involving topics and events that are widely reported in the media, we consult credible open source media sources (e.g., Reuters, BBC, AP) to determine the outcome of a question. Some questions may specify that the resolution will be determined by specific sources (e.g., UNHCR data, Moody's). When in doubt, please check the information provided in the “More Info” section of each question for guidance specific to that question.