Swift Centre

📋Page Status

Page Type:ContentStyle Guide →Standard knowledge base article

Quality:50 (Adequate)⚠️

Importance:35 (Reference)

Last edited:2026-02-01 (5 days ago)

Words:2.4k

Structure:

📊 2📈 0🔗 13📚 10•18%Score: 12/15

LLM Summary:Swift Centre is a UK forecasting organization that provides conditional forecasting services to various clients including some AI companies, but is not primarily focused on AI safety. While they demonstrate good forecasting methodology and track record, their relevance to AI risk is limited to client services rather than dedicated research or advocacy.

Issues (1):

QualityRated 50 but structure suggests 80 (underrated by 30 points)

Quick Assessment

Dimension	Rating	Notes
Focus	Forecasting	Conditional forecasting, scenario analysis, horizon scanning
Geography	UK	Registered in London, global forecaster network
Founding	2022	March 25, 2022 incorporation
Funding	$2M+	Future Fund grant (July 2022)
Team Size	Small	Core team of 3+ leaders, network of top 1% forecasters
Track Record	Strong	2025 Brier Score of 0.115 on public forecasts
AI Safety Relevance	Medium	Client work with AI organizations, not core mission

Key Links

Source	Link
Official Website	swiftcentre.org

Overview

The Swift Centre (full name: Swift Centre for Applied Forecasting) is a UK-based nonprofit organization that specializes in conditional forecasting, scenario analysis, and horizon scanning to help organizations make better decisions under uncertainty.¹ Founded in March 2022 by Michael Story, a former managing director and qualifying superforecaster from the Good Judgment Project, the organization bridges rigorous forecasting research with real-world strategic decision-making.²

Unlike traditional forecasting focused on general predictions, Swift Centre emphasizes conditional “what if?” questions tied to specific decisions, policies, or strategies. For example, rather than forecasting whether oil prices will rise, they might forecast: “If the US undertakes military action in Venezuela, will Chevron cease oil production/export there before 2027?”¹ This approach allows organizations to stress-test different scenarios and make more robust strategic choices.

The organization employs a global network of highly-calibrated forecasters ranked in the top 1% worldwide, including former US national security advisors and diplomats.² Their client portfolio spans AI safety organizations (Anthropic, Google DeepMind), government bodies (UK Department for Science, Innovation and Technology, Oxfordshire County Council), philanthropic funders (Open Philanthropy), and private sector clients (Financial Times, CapX).¹ As of their 2025 performance review, Swift Centre achieved an aggregate Brier Score of 0.115 on 13 resolved public forecasts, indicating strong calibration on low-probability, high-impact events.³

History

Founding and Early Development

Swift Centre was incorporated on March 25, 2022, as a private company limited by guarantee without share capital, with its registered office at 102 Fulham Palace Road, London, England.⁴ The organization was founded by Michael Story, who brought extensive forecasting credentials from his role as a qualifying superforecaster in the original Good Judgment Project and as a former managing director at Good Judgment Inc.² Story had previously applied forecasting, psychometrics, and risk quantification methodologies in hedge funds and consulting, bringing practical business experience alongside forecasting expertise.

The founding mission was to cut through “noise, uncertainty, and hype” by providing scenario analysis and horizon scanning that would help organizations anticipate future events rather than merely react to them.⁵ From the outset, Swift Centre differentiated itself by emphasizing conditional forecasting over general predictions, and by focusing on decision-relevant questions that could directly inform strategic choices.

Funding and Growth

In July 2022, just four months after incorporation, Swift Centre received $2 million in funding from the Future Fund.⁶ This substantial early-stage funding allowed the organization to build out its operations and begin publishing forecasts from expert forecasters and financial professionals, with a stated intention to focus on longtermist-relevant topics and trends from Our World in Data.⁶

The organization quickly established itself with high-profile clients across multiple sectors. Early engagements included work with the Open Nuclear Network, where Swift Centre facilitated a conference of technical experts on geopolitical nuclear risks, conducting forecasting principles training, calibration exercises, and software onboarding to produce estimates that informed multi-year risk responses.⁷ Sarah Laderman from Open Nuclear Network at One Earth Future Austria praised the team’s professionalism and their ability to help experts quickly absorb complex forecasting methodologies.⁷

Recent Developments (2024-2025)

By 2025, Swift Centre had developed a track record substantial enough to publish formal performance reviews. Their “2025 - A Year in Review” analyzed 13 public forecasts resolved during the year, achieving an aggregate Brier Score of 0.115—an excellent result that demonstrated strong calibration on low-probability, high-impact events.³ The review transparently acknowledged both successes and failures: they correctly predicted no major India-Pakistan war, no H5N1 pandemic, and resilient global oil markets despite the Israel-Iran conflict, but underestimated global LNG price growth due to an unanticipated decline in Chinese imports.³

The organization also expanded its public presence through a Strategic Insights hub offering timely public forecasts and expert analysis, along with an “Ahead of the Curve” subscription service.⁸ In April 2025, they released forecasts on UK ransomware risks, highlighting rising threats and the need for greater transparency in cybersecurity.⁹

Methodology and Services

Forecasting Approach

Swift Centre’s methodology draws from rigorous scientific research developed through structured forecasting tournaments, psychological studies, and statistical methodologies pioneered by projects like the Good Judgment Project.⁵ Their approach emphasizes several key principles:

Conditional forecasting: Rather than making unconditional predictions about the future, they focus on questions structured as “If X happens, what is the probability Y will occur?” This allows decision-makers to evaluate different scenarios and their likely consequences.
Calibration and validation: All forecasters in their network are measured, ranked, and publicly validated, ensuring that probability estimates are well-calibrated rather than merely confident-sounding guesses.¹
Collaborative elicitation: For complex questions requiring domain expertise, Swift Centre brings together their top forecasters with client-nominated subject matter experts to combine deep knowledge with forecasting skill.¹⁰
Transparent track records: The organization publishes performance reviews that acknowledge both accurate and inaccurate forecasts, demonstrating accountability and enabling continuous improvement.³

Service Offerings

Swift Centre provides fully managed forecasting services structured in four stages:¹⁰

Team training: Workshops teaching forecasting methodologies and building group forecasting efficacy, helping client teams develop internal capabilities.
Internal facilitation: Using client teams to forecast organization-specific topics through structured workshops, building buy-in while leveraging internal knowledge.
Collaborative estimates: Bringing together Swift Centre’s expert forecasters with client-nominated subject matter experts to produce rigorous, contextually-informed probability estimates.
Professional reports: Delivering polished reports with explicit rationales and analysis, along with ongoing monitoring and periodic updates as situations evolve.

Leadership and Team

Key Personnel

Michael Story (Founder and Director) leads the organization with a unique combination of forecasting excellence and practical business experience.² As a qualifying superforecaster in the original Good Judgment Project, Story demonstrated exceptional predictive accuracy. His previous role as Managing Director at Good Judgment Inc. gave him experience scaling forecasting methodologies for organizational clients. Before entering the forecasting field, he applied quantitative methods in hedge funds and consulting, and has a background as a startup founder with a Masters in Mathematics.²

James Newport serves as Executive Director, bringing government and policy expertise to complement the forecasting focus.² Newport was previously the AI & Digital Transformation Lead at Transport Canada and served as Global Risk & Scenario Forecaster at the Forecasting Research Institute, giving him deep experience in applying forecasting to policy questions.

Eleanor Parr serves as Chief of Operations, overseeing the operational frameworks, forecasting methodologies, project management, forecaster coordination, and publication workflows that ensure consistent quality and timely delivery.² Her role is critical to maintaining rigorous standards as the organization scales its client work.

Forecaster Network

Beyond the core leadership team, Swift Centre draws on a global network of highly-calibrated forecasters, all ranked in the top 1% worldwide through measured, publicly validated performance.¹ This network includes individuals with impressive credentials, such as a former National Security Special Advisor to the US Vice President and a former Senior US Diplomat to China, Hong Kong, and Taiwan.² This combination of forecasting skill and domain expertise allows the organization to tackle complex geopolitical and strategic questions.

Client Work and Impact

AI Safety and Governance

While AI safety is not Swift Centre’s primary mission, they have engaged with several prominent AI organizations on forecasting questions relevant to AI capabilities and governance. Their most significant AI-related project was with Google DeepMind, where they tasked top forecasters with AI experience to predict how future AI systems might perform on capability evaluations and to assess whether AI would rise in status as a global issue.⁷ The forecasts were integrated into a published paper from DeepMind, demonstrating how rigorous probabilistic forecasting can inform AI research and strategy.⁷

Swift Centre’s client list includes Anthropic and the Centre for the Governance of AI, suggesting ongoing work on AI-related questions.¹ However, the organization does not appear to have dedicated research programs or initiatives specifically focused on AI alignment, AI safety, or existential risk from AI. Their work in this domain is primarily client-driven rather than mission-focused.

Other Domains

Beyond AI, Swift Centre works across diverse sectors:

Nuclear security: Their engagement with the Open Nuclear Network involved facilitating expert workshops on geopolitical nuclear risks, helping technical experts produce calibrated forecasts that informed multi-year organizational responses.⁷

Business strategy: Client testimonials highlight Swift Centre’s ability to translate forecasts into actionable business insights. A representative from Metrea Aerospace noted: “We worked with the team from Swift Centre on a project to forecast risks to one of our biggest revenue drivers. It was an absolute pleasure… they’re professional, responsive and practical, working hard to understand not just the percentage probabilities… but what this would actually mean for our business.”¹⁰

Government and policy: Engagements with the UK Department for Science, Innovation and Technology and Oxfordshire County Council demonstrate Swift Centre’s role in supporting public sector decision-making under uncertainty.¹

Media and analysis: Partnerships with the Financial Times and CapX suggest work on forecasting questions relevant to journalism and public commentary.¹

Track Record

Swift Centre’s 2025 performance review provides concrete evidence of forecasting accuracy.³ On 13 resolved public forecasts, they achieved an aggregate Brier Score of 0.115 (where 0 is perfect and 0.25 is the expected score for random 50-50 guesses). Notable successes included:

Correctly forecasting no major India-Pakistan war during the period
Accurately predicting that H5N1 would not develop into a pandemic
Forecasting resilient global oil markets despite ongoing Israel-Iran conflict
Identifying US-Venezuela risks ahead of mainstream media coverage

The organization also demonstrated intellectual honesty by acknowledging forecasting failures, particularly underestimating global LNG price growth due to unanticipated changes in Chinese import patterns.³ This transparent approach to performance review strengthens confidence in their methodology even as it acknowledges specific predictive failures.

Relationship to Effective Altruism and Rationality Communities

Swift Centre has connections to the Effective Altruism community, evidenced by funding from the Future Fund and clients including Open Philanthropy.⁶¹ The organization has a dedicated topic page on the EA Forum, indicating relevance to that community’s interests.⁶

However, unlike organizations like MIRI or Redwood Research that emerged directly from rationalist/EA communities with explicit AI safety missions, Swift Centre positions itself as a professional forecasting service organization. Its leadership comes from forecasting and policy backgrounds rather than the rationalist community, and its work spans commercial, governmental, and nonprofit sectors beyond EA priorities.

The organization’s methodology draws on research popularized in the rationality community—particularly work from the Good Judgment Project on calibration, Brier scores, and superforecasting—but applies these methods as professional services rather than as community-building or movement-focused activities.

Limitations and Criticisms

Self-Acknowledged Forecasting Challenges

Swift Centre has demonstrated intellectual honesty by publicly acknowledging forecasting failures in its performance reviews. In the 2025 review, the organization noted its failure to anticipate persistent Asian price sensitivity and the decline in Chinese LNG imports that affected global growth patterns.³ The forecasting model had assumed 3.7% trend growth following 2024’s 2.8% rebound, but actual patterns diverged due to factors not fully captured in the analysis.

This transparency is valuable for calibrating expectations about forecasting limitations—even excellent forecasters working with rigorous methodologies will miss important developments, particularly when facing novel situations or structural changes in the systems being forecast.

Limited Public Track Record

While Swift Centre has published performance data for 2025, the organization’s relatively short operational history (founded 2022) means there is limited longitudinal data on forecasting accuracy across different types of questions and time horizons. The 13 resolved forecasts from 2025 provide valuable signal, but a longer track record would enable more confident assessment of their capabilities.

Scope Limitations

Swift Centre specializes in probabilistic forecasting but does not claim to provide comprehensive strategic advisory services, deep subject-matter expertise, or implementation support. Organizations using their services still need to interpret forecasts in context, integrate them with other sources of information, and translate probabilistic estimates into strategic choices. The testimonial from Metrea Aerospace highlighting Swift Centre’s effort to explain “what this would actually mean for our business” suggests this translation is not always straightforward.¹⁰

Key Uncertainties

Several important questions about Swift Centre remain difficult to answer based on publicly available information:

Sustained forecasting accuracy: Will Swift Centre maintain its 2025 performance levels (Brier Score 0.115) as it scales and tackles increasingly diverse questions? Forecasting tournaments suggest accuracy can degrade when forecasters work on topics outside their expertise or when questions become too complex.
Organizational scaling: Can Swift Centre preserve its rigorous methodology and forecaster quality as it grows? Many professional services organizations face quality-consistency challenges when expanding beyond a small core team.
Impact on decision-making: While client testimonials are positive, it remains unclear how frequently Swift Centre’s forecasts materially change organizational strategies compared to providing confirmation or marginal adjustments to existing plans. Measuring the counterfactual impact of forecasting services is inherently difficult.
Long-term financial sustainability: The $2 million Future Fund grant provided crucial early-stage funding, but the organization’s long-term revenue model through client services is not publicly documented. Whether Swift Centre can sustain operations through fee-for-service work or will require additional philanthropic funding remains uncertain.
AI safety contribution: Given that several prominent AI safety organizations are clients, what specific forecasting questions in AI safety benefit most from Swift Centre’s approach? Conditional forecasting seems well-suited to questions like “If AI system X is deployed, what is the probability of outcome Y?” but may be less useful for deeply uncertain questions about novel AI capabilities or alignment techniques.