Skip to content
@forecastingresearch

Forecasting Research Institute

Advancing the science of forecasting for the public good

We develop forecasting methods to improve decision-making on high-stakes issues. First-generation forecasting research—spearheaded by FRI Chief Scientist Philip Tetlock and coauthors—focused on establishing a rigorous standard for prediction accuracy. The next generation of work aims to channel this approach into real-world relevance. We work with policymakers and nonprofit organizations to design practical forecasting tools, and test them in large experiments.

Our research centers on:

  1. Producing high-quality forecasting questions about complex, long-run topics;
  2. Novel methods for resolving unresolvable questions;
  3. Testing the robustness of forecasting techniques across different domains and contexts; and
  4. Using forecasting tools to help organizations make better decisions.

Pinned Loading

  1. xpt-lib Public

    Provides the data (anonymized) and code necessary to replicate the results of the Existential Risk Persuasion Tournament (XPT). The XPT was a forecasting tournament that explored potential threats …

    R 5 1

  2. aggutils Public

    Utilities for Aggregating Probabilistic Forecasts

    R 3 1

  3. forecastbench Public

    A dynamic forecasting benchmark for LLMs

    Python 17 1

  4. forecastbench-datasets Public

    Forecastbench Datasets, updated nightly

    JSON 9

  5. adversarial-collab Public

    Provides the data (anonymized) and code necessary to replicate the results in FRI's report "Roots of Disagreement on AI Risk: Exploring the Potential and Pitfalls of Adversarial Collaboration."

  6. voivod Public

    Provides a set of statistics for evaluating forecasting questions (not forecasts, but the questions themselves, with respect to another question), i.e. assessing the cruxiness of questions. Current…

    R

Repositories

Showing 9 of 9 repositories
  • forecastbench-datasets Public

    Forecastbench Datasets, updated nightly

    JSON 9 CC-BY-SA-4.0 0 0 0 Updated Mar 30, 2025
  • forecastbench Public

    A dynamic forecasting benchmark for LLMs

    Python 17 MIT 1 18 2 Updated Mar 28, 2025
  • adversarial-collab Public

    Provides the data (anonymized) and code necessary to replicate the results in FRI's report "Roots of Disagreement on AI Risk: Exploring the Potential and Pitfalls of Adversarial Collaboration."

    0 MIT 0 0 0 Updated Feb 5, 2025
  • xpt-lib Public

    Provides the data (anonymized) and code necessary to replicate the results of the Existential Risk Persuasion Tournament (XPT). The XPT was a forecasting tournament that explored potential threats to humanity, with a focus on artificial intelligence, biosecurity, climate, and nuclear arms.

    R 5 1 0 0 Updated Feb 5, 2025
  • xpt-ijf-replication Public

    A repository containing code needed for replicating results from "Subjective-Probability Forecasts of Existential Risk: Initial Results from a Hybrid Persuasion-Forecasting Tournament"

    R 0 0 0 0 Updated Dec 4, 2024
  • utils Public

    Utility files (mainly GCP functions)

    Python 0 MIT 0 0 0 Updated Oct 10, 2024
  • .github Public
    0 MIT 0 0 0 Updated Oct 1, 2024
  • voivod Public

    Provides a set of statistics for evaluating forecasting questions (not forecasts, but the questions themselves, with respect to another question), i.e. assessing the cruxiness of questions. Currently the package includes several implementations each of value of information (VOI) and value of discrimination (VOD).

    R 0 MIT 0 2 0 Updated Jul 11, 2024
  • aggutils Public

    Utilities for Aggregating Probabilistic Forecasts

    R 3 1 1 0 Updated Feb 9, 2024