Good Judgment Open marketing: how a crowd-forecasting site uses partner challenges, leaderboards and scoring in plain words

How Good Judgment Open markets crowd forecasting with challenges for The Economist and Harvard Kennedy School, leaderboards, Brier scores and badges.

The Convrt team13 min read17 sources
2015year the public tournament began
11,410forecasters in Foxes Ask
22challenges listed
The short version

Good Judgment Open is the public crowd-forecasting site run by Good Judgment, the firm co-founded by Philip Tetlock. Its pitch fits in one line: Join the internet's smartest crowd.

The product is the challenge: a themed set of questions with its own leaderboard. Partners lend their names, from The Economist, in a 12th annual challenge, to a course at Harvard Kennedy School, and public data such as FRED and S&P Global PMIs.

Around the challenges sit three accuracy scores explained in plain words, a crowd forecast on every question page, 14 kinds of badge and one line that sets it apart: you can share your reasoning with other forecasters.

Good Judgment Open is a forecasting site that grows through challenges rather than ads. This breakdown covers who GJ Open is, the research project behind it, challenges as the product, the challenge with The Economist, the Harvard Kennedy School course, where the crowd is, where the forecasts are, public data as a challenge, one index as 57 questions, a challenge for every beat, what the questions cover, scoring in plain words, scoring against the crowd, the crowd on one question, badges that reward reasoning, the rules that keep scores honest, the path to Superforecaster and what any prediction market brand can copy.

Good Judgment Open at a glance
  1. 012015Public tournament began, in September.
  2. 0222Challenges on the Challenges page.
  3. 0311,410Forecasters in Foxes Ask.
  4. 04189,003Forecasts in Foxes Ask.
  5. 053Accuracy scores on every profile.
  6. 0614Kinds of badge.
Good Judgment Open at a glance, from its Challenges page, FAQ and Wikipedia (read 19 Nov 2026).
How a forecaster moves through GJ Open
→ 01 Challenge Pick a theme with its own leaderboard
→ 02 Question Read how it will be resolved
→ 03 Forecast Give a probability
→ 04 Rationale Share the reasoning
→ 05 Update Change it as news lands
→ 06 Score Relative Brier Score against the crowd
How a forecaster moves through GJ Open, from its FAQ and question pages (read 19 Nov 2026).

“Unlike prediction markets and other forecasting sites, you can share your reasoning with other forecasters to challenge your assumptions.”Good Judgment Open, FAQ (read Nov 2026)

Born out of a research project

The FAQ gives the origin in one line: GJ Open was born out of the Good Judgment Project. Wikipedia dates the project: The Good Judgment Project began in July 2011. Its team recruited forecasters and gave them basic tutorials on forecasting best practice and overcoming cognitive biases.

Starting in September 2015, Good Judgment Inc has been running a public forecasting tournament at the Good Judgment Open site. The scoring carried over too: Predictions are scored using Brier scores.

That history is the brand. The site does not need to claim expertise when its founder wrote the book, and every page points to Superforecasting: The Art and Science of Prediction as the reading list.

The story behind the site
→ 01 July 2011 The Good Judgment Project begins
→ 02 Training Tutorials on best practice and biases
→ 03 Scoring Every forecast gets a Brier score
→ 04 September 2015 The GJ Open tournament starts
→ 05 Challenges Themed sets with partners
→ 06 Today 22 challenges listed
The story behind the site, from Wikipedia and the GJ Open FAQ (read 19 Nov 2026).

Challenges are the product

GJ Open packages its questions as competitions. Competitions on GJ Open are called challenges. Challenges are collections of questions organized by a theme or topic. Each challenge has its own leaderboard, which ranks forecasters by comparative accuracy against the forecasters participating in a specific challenge.

That one design choice does most of the marketing. A challenge has a name, a theme, an end date and a winner, so it can be co-branded, announced and run again next year. The home page tells visitors Be sure to check out all of our active challenges, our featured questions, and our unfiltered list of all open forecasting questions.

Sponsors get named on the front page: UBS Asset Management, The Economist and Harvard Kennedy School are listed among them. On 19 November 2026 the Challenges page listed 22 challenges, from awards season to University of Mannheim: Champions League Leadership Forecasting Challenge, a challenge on clubs that several clubs enter the 2026/27 UEFA Champions League under freshly appointed head coaches, with 31 forecasters.

Challenges built with partners and data
  1. 01The EconomistThe World Ahead, 12th year.
  2. 02Harvard Kennedy SchoolA course challenge.
  3. 03UBS Asset ManagementNamed as a sponsor.
  4. 04University of MannheimChampions League coaches.
  5. 05FREDEvery question sourced from it.
  6. 06S&P GlobalMonthly PMI questions.
Challenges built with partners and public data, from the GJ Open home and Challenges pages (read 19 Nov 2026).
The Economist: The World Ahead 2027
  1. 0112thAnnual collaboration.
  2. 0225Questions planned.
  3. 0322Needed to be eligible to win.
  4. 04Late 2027Winner published in The Economist.
  5. 05Top 20%Earn the Economist MVP badge.
  6. 06Top 5Badges for each year.
The Economist: The World Ahead 2027 challenge, from its challenge page and the FAQ (read 19 Nov 2026).
The Harvard Kennedy School challenge
  1. 01FifthCollaboration with the school.
  2. 02SemesterForecasts over one term.
  3. 03CourseworkStudents assigned the questions.
  4. 04PublicAnyone can join in.
  5. 05281Forecasters.
  6. 063,797Forecasts.
The Harvard Kennedy School challenge, from its challenge page and the Challenges page (read 19 Nov 2026).

Where the crowd is

The biggest challenges are the ones the crowd and the workshops feed. Foxes Ask asks people to forecast on topics suggested by Good Judgment Open forecasters through our crowd discussion question, during Office Hours, or through our Help Desk, and it had 11,410 forecasters. We look forward to seeing your suggested questions here!

The Superforecasting Workshops challenge has people forecast on questions generated by participants of Good Judgment's public and private Superforecasting Workshops. It had 8,399 forecasters, so the workshops and the site feed each other.

Among the themed challenges, In the News 2026 had 3,558 forecasters and AI in 2026 and Beyond had 840. The 2026-27 Sportsball challenge had 621, A Challenge of Interest! had 277 and the Red Carpet challenge had 53.

Forecasters per challenge
Foxes Ask 11,410
Superforecasting Workshops 8,399
In the News 2026 3,558
AI in 2026 and Beyond 840
2026-27 Sportsball 621
Harvard Kennedy School 281
A Challenge of Interest! 277
The Economist 2027 62
Forecasters per challenge, from the Challenges page (read 19 Nov 2026).

Where the forecasts are

Forecasts tell a different story from sign-ups, because people return to update. Foxes Ask had 189,003 forecasts, In the News 2026 had 136,963 and the workshops challenge had 110,356.

The FAQ pushes that habit. You are strongly encouraged to update your forecasts frequently no matter how wrong you think previous forecasts might be. A forecast also stays live on its own: we carry that forecast forward each day until you update it by submitting a new forecast.

Forecasts per challenge
Foxes Ask 189,003
In the News 2026 136,963
Superforecasting Workshops 110,356
2026-27 Sportsball 10,844
AI in 2026 and Beyond 9,054
A Challenge of Interest! 8,359
Harvard Kennedy School 3,797
FRED Q4 2026 1,591
Forecasts per challenge, from the Challenges page (read 19 Nov 2026).
The FRED challenge
  1. 011991FRED launched by the St. Louis Fed.
  2. 02845,000+Data series in FRED.
  3. 0314Questions, all sourced from FRED.
  4. 0412 of 14Needed to qualify for the badge.
  5. 05Top 3Earn a special badge.
  6. 06InterviewFor a first-time winner.
The FRED challenge, from its challenge page and the Challenges page (read 19 Nov 2026).
The Purchasing Managers’ Indices challenge
  1. 0157Active questions.
  2. 0248Forecasters.
  3. 031,554Forecasts.
  4. 04Late 1940sWhen PMIs were first created.
  5. 05MonthlyHow often S&P Global publishes them.
  6. 06Over 50Signals expansion.
The Purchasing Managers’ Indices challenge, from its challenge page, a question page and the Challenges page (read 19 Nov 2026).
A challenge for every beat
→ 01 Awards Emmys, Grammys and Oscars
→ 02 Sport The Sportsball challenge
→ 03 Rates A Challenge of Interest
→ 04 AI AI in 2026 and Beyond
→ 05 Football Champions League head coaches
→ 06 News New questions every week
A challenge for every beat, from the challenge pages (read 19 Nov 2026).

What the questions cover

In the News 2026 shows how wide the site goes. Its tag filter lists Business (157), Economic Indicators (91), Technology (46) and Finance (45), then Society (24), Environment (16) and Sports (14) and Health (7) and Entertainment (5), among others.

Topic tags work as navigation. A forecaster who likes markets can filter to finance, and one who likes awards can stay with entertainment, without leaving the challenge.

Questions by tag, In the News 2026
Business 157
Economic Indicators 91
Technology 46
Finance 45
Society 24
Environment 16
Sports 14
Health 7
Entertainment 5
Questions by tag in In the News 2026, selected tags, from its challenge page (read 19 Nov 2026).
The Brier score in plain words
  1. 010The best possible score.
  2. 022The worst possible score.
  3. 030.18A 70% yes that happened.
  4. 04GolfLower is better.
  5. 05DailyA score for every day.
  6. 06WeatherWhere the score began.
The Brier score in plain words, from the FAQ (read 19 Nov 2026).
How the Relative Brier Score is built
→ 01 Daily A Brier score for each day
→ 02 Average Across the life of the question
→ 03 Crowd Minus the crowd median
→ 04 Participation Times the share of days active
→ 05 Challenge Added up across its questions
→ 06 Leaderboard Below zero beats the crowd
How the Relative Brier Score is built, from the FAQ (read 19 Nov 2026).

The crowd on one question

Every question page shows what the crowd thinks. The page on the UK Services PMI for February 2027 lists answer ranges with the columns Possible Answer Crowd Forecast Change in last 24 hours Change in last week Change in last month.

On 19 November 2026 the top range read at least 52.0, but less than 54.0, at 55.40%, with 21.70% on 50.0 to 52.0. The page gives the last reading for context: For September 2026, S&P Global reported UK Services PMI as 52.1. Reported "flash" estimates are immaterial.

The page ends with a pointer to help: Confused? Check our FAQ or ask us for help.

Crowd forecast, UK Services PMI, February 2027
Less than 46.0 1.0%
46.0 to 48.0 1.7%
48.0 to 50.0 6.3%
50.0 to 52.0 21.7%
52.0 to 54.0 55.4%
54.0 to 56.0 10.9%
56.0 or more 3.0%
Crowd forecast in percent for the UK Services PMI, February 2027, from the question page (read 19 Nov 2026).

Badges that reward the reasoning

The FAQ names what sets GJ Open apart: Unlike prediction markets and other forecasting sites, you can share your reasoning with other forecasters to challenge your assumptions. The badges are built to reward exactly that.

Badges are public awards you can earn by participating in GJ Open. We award new badges (and expire old ones) on the first of each month. Badges do not affect how we score your forecasts.

The Influencer badge says it best: The best forecasters don't just put their reputation on the line, but explain how they chose their probabilities. It needs at least 2 comments (as rationales accompanying forecasts or replies to other forecasters) with at least 2 upvotes each. Rationale badges start with Show your work! You'll earn a badge for a total of 25, 50, and 100 rationales.

Others reward habits. For Frequent Forecaster, You will earn this badge if you made at least 10 forecasts in the previous calendar month., and if you want to keep it, you have to stay active! The rest climb in levels. You earn this permanent badge by forecasting on 50, 100, 250, or 500 cumulative questions. You earn this permanent badge by updating your forecasts 100, 200, 500, or 1000 times. You earn this permanent badge by receiving 50, 100, 250, 500, or 1000 cumulative upvotes. You earn this permanent badge by posting 50, 100, 250, 500, or 1000 rationales with links.

The top tier is about accuracy: You earn this persistent badge by being in the top 10 in 25, 50, 100, or 250 questions, and since June 2022 it also needs a score better than the crowd. Forecasters pick what to show, since you can choose which badges you want to feature most prominently on your profile.

Badges for activity and reasoning
  1. 01Frequent Forecaster10 forecasts last month.
  2. 02Influencer2 comments with 2 upvotes each.
  3. 03Rationale25, 50 and 100 rationales.
  4. 04ResearcherRationales with links.
  5. 05Updater100 to 1,000 updates.
  6. 06Top ForecasterTop 10, better than the crowd.
Badges for activity and reasoning, from the FAQ (read 19 Nov 2026).

Rules that keep the score honest

Trust in a leaderboard comes from rules people can read. There is no way to withdraw from a question or delete a forecast. Data questions use first prints: forecasting questions will be resolved when relevant data are initially released, and subsequent revisions to the data will not affect the question’s resolution. Discussion is kept apart: Open questions are never scored for accuracy.

The FAQ also admits the limits of its own rule. The usual Brier score does not give any credit for “near-misses.”, so ordered questions use a special rule. And a question that outlives its challenge is not wasted: GJ Open will roll the question over into the next version of the challenge or a similar challenge.

The etiquette policy fits on one line: “Be kind, respect others, and stay on topic.” The goal is our goal is to provide a fun, respectful, and constructive place for everyone to do so. When an argument stalls, the advice is to let your forecast speak for itself. Eventually, scores will show who was right and who was wrong.

Help starts with the community too: First, consult the GJ Open crowd. Other forecasters on the site are your best resource. And the crowd shapes what comes next: To suggest a question, you can post your suggestion on the webpage itself.

The path to Superforecaster

The site keeps one goal in front of every visitor. The FAQ answers how to become a Superforecaster in one line: In short: elite accuracy, along with quality comments and collegiality.

The advice to get there is reading and practice: the book Superforecasting: The Art and Science of Prediction by Philip E. Tetlock and Dan Gardner, and then Most importantly, PRACTICE on questions here on GJ Open!

What any prediction market brand can copy

What any prediction market brand can copy
→ 01 Challenges Group questions by theme
→ 02 Partners Co-brand with media and schools
→ 03 Leaderboards One for each challenge
→ 04 Scores Explain them in plain words
→ 05 Reasons Reward written rationales
→ 06 Crowd Let users suggest questions

GJ Open markets forecasting as a game with rules, seasons and partners. It packages questions as named challenges, gives each one a leaderboard and lends the format to a magazine, a school and public data sets.

The lessons are simple. Run themed seasons with a clear winner. Co-brand them with media and schools. Explain the score in plain words. Reward people who write down their reasoning. Then cut the clearest minute of every challenge launch and every winner story into clips and post them daily on TikTok, Instagram and YouTube Shorts from accounts built for short form.

How we would run it for you Creator campaigns for sportsbooks

We run creator and clipping campaigns for sportsbooks and prediction markets, with every post checked against the ad rules in each state and tracked for views, link clicks and click-through rate. The system in this article, run for you.

Frequently asked questions

GJ Open is a crowd-forecasting site where people make probability forecasts on future events, see how accurate they were and compare their accuracy with the crowd. It is owned and operated by Good Judgment, a forecasting services firm, and was born out of the Good Judgment Project.

No. Its FAQ sets it apart in one line: unlike prediction markets and other forecasting sites, you can share your reasoning with other forecasters to challenge your assumptions. Forecasters give probabilities and are ranked by accuracy, not by trading.

It reports three numbers: the Brier Score, the Median Score and the Relative Brier Score. Lower is better, like in golf. The Relative Brier Score compares a forecaster with the crowd median each day, scaled by how many days they had a forecast, and it ranks every challenge leaderboard.

On 19 November 2026 its Challenges page listed The Economist: The World Ahead 2027, its 12th annual collaboration with The Economist, a Harvard Kennedy School course challenge, a University of Mannheim Champions League challenge and challenges built on FRED and S&P Global PMI data.

Package questions as themed challenges, give each one its own leaderboard, co-brand challenges with media and schools, explain scoring in plain words, reward written reasoning with badges, and let the crowd suggest the next questions.

More breakdowns

All breakdowns →

Want distribution done properly?

Apply and we'll map your existing content to real reach on accounts you own.

Pinch to zoom · swipe down to close

Apply to work with us.

Five quick questions, then pick a time.