February 1, 2023 update: Metaculus has released the 'Forecasting Our World in Data: The Next 100 Years' report, detailing forecasters' predictions and analysis on the questions developed by Our World in Data for this tournament.

The Forecasting Our World in Data tournament probes the long-term future, delivering predictions on topics like global investment in AI, world life expectancy, CO2 emissions, and more on time horizons from one to 100 years. This close collaboration with Our World In Data (OWID) collects predictions across 30 different measures of technological advancement, global development, and social progress.

Overview

OWID is one of the world’s largest open online repositories of data, and it's an incredible source of information about both humanity’s progress and challenges. Your forecasts will be used to generate likelihoods of specific outcomes in humanity’s near- and long-term future. And throughout the tournament, a team of Metaculus Pro Forecasters will also predict alongside the public tournament.

A $20,000 prize pool will reward accurate forecasts on 1 and 3-year outcomes and insightful comments on 10, 30, and 100-year outcomes.

The release of new questions is staggered, with 10 questions released on each of the following dates:

Wednesday, October 12
Wednesday, October 19
Wednesday, October 26

We want to extend our thanks to Future Fund for the grant funding making this project possible. And as always, we are incredibly grateful to the Metaculus community for your forecasts contributing to a greater understanding of humanity’s near- and long-term future!

Read on for additional detail about how this tournament is structured, how prizes are awarded, and how forecasts are scored.

Tournament Prize Structure

Scoring Overview

The Forecasting Our World in Data Tournament consists of three separate prize pools totaling $20,000. The tournament will feature approximately 30 question groups, with each group asking about time horizons of 1, 3, 10, 30, and 100 years.

There will be separate leaderboards for the questions asking about 2023 and for those asking about 2025, with an $8,000 prize pool for each of these time horizons. Prizes for each time horizon are entirely determined by the questions asked for that time horizon. The questions which have resolved by the cutoff dates shown below will be scored for prizes.

Note: We may move the date prizes are awarded to precede the cutoff date if we can determine that all of the questions that will resolve before that date have indeed been resolved.

2023 questions: $8,000 awarded by May 1, 2025. Questions resolving after the cutoff date will not be scored for prizes.
2025 questions: $8,000 awarded by May 1, 2027. Questions resolving after the cutoff date will not be scored for prizes.
No prizes are awarded for scores on the 2032, 2052, and 2122 questions. However there is a commenting prize as described below.

Questions are set to close on December 31 of the year in question. Read more about expected resolution dates in the "Expected Resolution" section below.

Commenting Prize

In addition to the prizes for the questions asking about 2023 and 2025 there will be a commenting prize pool of $4,000 for the questions asking about 2032, 2052, and 2122. There will be 19 commenting prizes awarded for comments made before January 1, 2024. In January of 2024, three Metaculus Admins will independently read through all of the comments on all of the question groups in the Forecasting Our World in Data tournament and assign a rating to each comment out of 10. Ratings will be aggregated, and prizes will be awarded to the top 19 comments as shown below:

$500 each for the top three comments
$250 each for the fourth place through ninth place comments
$100 each for the tenth place through 19th place comments

Individual users may receive multiple comment prizes up to a maximum prize award of $1,500. The comments must pertain to either the 2032, 2052 or 2122 questions to be eligible (or any combination of those periods). Comments may also pertain to the 2023 and 2025 questions, though portions which are relevant exclusively to those periods will be ignored for the purpose of awarding prizes. The comments will be judged based on how they meet some or all of the criteria described below:

The comment pertains to the 2032, 2052, or 2122 time horizon (or some combination of these)
The comment is well-reasoned and clearly written
The comment provides valuable information and reasoning
The comment identifies mistakes in the community’s reasoning or clearly explains why the author’s forecast differs from the community’s
The comment describes connections between forecasts on different question groups, or inconsistencies between forecasts on different question groups

Tournament Scoring

The 2023 questions will be hidden for one month and the 2025 questions will be hidden for three months, during which the community median will not be visible. Questions for 2032, 2052, and 2122 will also be hidden for three months, however these time horizons will not be scored.

For the leaderboard, 25% of the coverage weight will be determined by participation during this hidden period and the other 75% will be based on participation during the visible period. Tournament scoring aims to balance rewarding independent forecasting skill (during the hidden period) and providing latecomers who miss the hidden period a realistic chance to finish in first place.

Expected Resolution

Some of the OWID metrics rely on underlying sources that have a long interval between the end of a given year and the updated data being published. A few others have update schedules that vary quite a bit. We think many of these are important questions worth asking, so we have included them in the tournament. Some of these questions may not resolve in time to be scored when prizes are awarded. Any questions which have not resolved by the cutoff date will be excluded from the tournament leaderboard, and will not be included when calculating coverage.

Currently we think that four of the questions will likely resolve after the cutoff dates. These four questions are listed below. Others could end up delayed and resolving after the cutoff dates as well. If these questions do in fact resolve after the cutoff dates then they will be excluded from the scoring used to award prizes and they will also be excluded from the coverage used to award prizes. Some of these questions will be launched on October 19 and October 26, and you will not have access to them until they have launched.

Here is a document containing notes on the underlying sources and the estimated resolution dates for each of the questions.

End Date and Leaderboard Functionality

This project features separate prizes for the questions which resolve in 2023 and 2025. Currently our tournament functionality only allows us to set one end date for the project, so we have set this date to January 1, 2024, after the 2023 questions will close on December 31, 2023. We intend to implement improvements to our tournaments to be able to show separate leaderboards for the different periods and to convey more information about end dates prior to prizes being awarded. Despite the tournament end date shown, the 2025 questions will be scored and prizes awarded as described above.

Dealing With Extinction

For the Forecasting Our World in Data tournament all questions resolve as Ambiguous in the event that humanity goes extinct or ceases to have a developed society (see the fine print in the questions for specifics) except for the following questions:

These questions (which launch on October 19) resolve below the lower boundary if total population or GDP per capita are below the lower boundaries for the relevant question. This allows us to separately collect estimates of the risk of human extinction and economic collapse for the time horizons of interest, while differentiating between the risk of extinction/collapse and the possibility of a substantial reduction in certain human activities, for example global CO2 emissions.