Global Risk

FoxCast Forecast Ledger: The First Public Scorecard

A short public record of forecast questions, probabilities, deadlines, and scored outcomes.

Published 2026-05-28 · 3 min · For: FoxCast readers, operators, buyers, and strategy teams.

Most market commentary disappears after it is written.

This ledger is built to do the opposite. If a forecast has a probability, it should also have a deadline and a rule for what counts. When the outcome is known, it should be scored.

That does not mean every useful piece of analysis belongs in a scorecard. A supply-chain map can explain a bottleneck. A foresight note can describe a product category taking shape. A field-level agriculture post can show what pressure is forming. The ledger is narrower. It is for questions that can be checked later.

Active Forecasts

Agriculture: pasture stress

Question: Will U.S. pasture and range rated poor plus very poor reach at least 35% for two consecutive USDA Crop Progress reports before August 31?

• Current probability: 46%
• Information cutoff: 2026-05-21
• Deadline: 2026-08-31
• Status: Open

Agriculture: wheat production revision

Question: Will USDA lower its 2026/27 U.S. all-wheat production estimate by at least 3% from the May 2026 WASDE level before September 30?

• Current probability: 42%
• Information cutoff: 2026-05-21
• Deadline: 2026-09-30
• Status: Open

Agriculture / Global Risk: diesel cost pressure

Question: Will EIA U.S. on-highway diesel average at least $6.00 per gallon for four consecutive weekly reports before November 30?

• Current probability: 36%
• Information cutoff: 2026-05-21
• Deadline: 2026-11-30
• Status: Open

Global Risk / Macro: long-yield persistence

Question: Will the U.S. 10-year Treasury yield trade at least 25 basis points above its May 14 close for five consecutive trading days before July 31?

• Current probability: 37%
• Information cutoff: 2026-05-26
• Deadline: 2026-07-31
• Status: Open

Critical Minerals / Global Risk: rare-earth licensing posture

Question: Will U.S.-China rare-earth export-control or licensing posture materially change before November 10?

• Current probability: 31%
• Information cutoff: 2026-05-28
• Deadline: 2026-11-10
• Status: Open

Critical Minerals: non-China graphite qualification

Question: Will a named non-China graphite or anode supplier announce customer qualification or repeat commercial orders before year-end?

• Current probability: 28%
• Information cutoff: 2026-05-21
• Deadline: 2026-12-31
• Status: Open

What Changed

The main change is simple: these questions are now in public view.

That matters because the useful question is usually not "is this important?" It is more specific: what would have to happen, by what date, for this pressure to become real enough to change a decision?

For agriculture, the cleanest current questions are tied to pasture, wheat, and diesel. Those are not abstract macro themes. They affect feed planning, acreage assumptions, freight, harvest costs, and cash-flow pressure. If the thresholds are crossed, the result should be visible in public data, not just in a narrative.

For critical minerals, the better question is not whether China is important. That is already obvious. The better question is whether buyer-visible licensing, qualification, or order behavior changes. Capacity announcements are useful background. Customer qualification and repeat orders are stronger signals.

Resolved Forecasts

Global Risk: China purchase commitments

Forecast: China announces dated U.S. energy, farm-goods, aircraft, or industrial-purchase commitments within 30 days of the Trump-Xi Beijing meeting.

• Probability: 43%
• Outcome: Yes
• Brier score: 0.3249

Lesson: The useful distinction was dated commitments versus general cooperation language. A cautious forecast can still resolve yes when the artifact lands.

How To Read The Score

A Brier score grades a probability forecast after the outcome is known. Lower is better.

If a 70% forecast happens, that is a good result. If a 90% forecast does not happen, the miss should hurt. The point is not to sound certain. The point is to make uncertainty specific enough that it can be judged later.

Unresolved forecasts do not get Brier scores yet. They get a probability, a deadline, and an outcome rule. Once the event resolves, the record gets updated.

The public ledger will stay free. It is the accountability layer. Deeper memos can explain the pressure maps, watchlists, and what would change the view, but the basic record should be visible.