The Housing Tribunal · Scoreboard

How often is our AI actually right?

Every forecast the Tribunal makes is listed here before the answer is known, and graded by a script when Realtor.com publishes the number. Nothing is removed, and no AI grades a forecast.

The Tribunal’s record

—of graded forecasts right (0 of 0)
—Brier score (0 is perfect; 0.25 is a coin flip)
3forecasts awaiting data
0void (the number didn’t move)

No forecast has come due yet. The first grades arrive when Realtor.com publishes December 2026, usually in the first days of the following month.

Every forecast

SittingStateThe callConfidenceOutcomeBrier
November 2026 California active listings for December 2026 higher than 56,016 80% pending
November 2026 Ohio active listings for December 2026 higher than 19,811 88% pending
November 2026 Texas median list price for December 2026 lower than $350,000 65% pending

Beat the Tribunal: players

Anyone can call the same questions on the docket. Same data, same grading script.

The leaderboard fills in when the first forecasts are graded. Make your calls on the docket before the sitting closes.

The Rate Call: players

The weekly game: will next week’s 30-year mortgage rate be higher or lower? Play this week’s round.

No graded rounds yet. Rounds are graded every Friday.

You are playing as: …

There are no accounts, names or emails. Your browser holds a random player code, and your handle is generated from it, so there’s nothing to sign up for and nothing for us to moderate.

How grading works: a forecast or call says a number will be higher or lower than a stated baseline. When the publisher releases the number, a script compares the two. Equal means void for the AI and every player. The script and the data are public: see the code.