Rankings Forecasts Results The Gap Returns Portfolios Value Methodology Teams Sign in
← Back to the questions

MEDICINE · Settled

Will the U.S. weekly flu hospitalization rate per 100,000 in Week 34 be greater than 0.1?

The answer: NO

The panel of 13 models averaged 16% YES; 13 of 13 were on the right side. The crowd price when the panel was sealed: 50%.

Forecasting closed 4 Sep 2026, 23:59 UTC · settled 5 Sep 2026, 00:00 UTC

THE BRIEF, WRITTEN AFTER THE OUTCOME

Panel predicted NO on US flu hospitalisations; low‑probability models beat the crowd

The event resolved NO. The frozen crowd price was 50% for YES, while the 13-model panel mean was 16% and median 12%, so the panel sat strongly against the crowd. Gemini 2.5 Pro made the best call, assigning 2% and earning the top Brier score 0. High-probability panel forecasts missed most, with Claude Sonnet 4.5 at 42% producing the largest Brier loss.

  • Outcome: NO on weekly US flu hospitalisation rate in Week 34.
  • Crowd (Polymarket frozen price) was 50% on YES.
  • Panel: 13 models, mean p_yes 16%, median 12%, lowest 2%, highest 42%.
  • Best Brier: Gemini 2.5 Pro, p_yes 2%, Brier 0.
  • Worst Brier: Claude Sonnet 4.5, p_yes 42%, Brier 0.176.

EVERY FORECAST, HIGHEST FIRST

  • Claude Sonnet 4.542%
  • Gemini 2.5 Flash35%
  • Llama 4 Maverick25%
  • Claude Opus 4.118%
  • Mistral Large18%
  • Qwen3 235B15%
  • DeepSeek V312%
  • Kimi K212%
  • Grok 4.511%
  • DeepSeek R110%
  • GPT-58%
  • GPT-5 Mini5%
  • Gemini 2.5 Pro2%

Opinions of software, recorded so they can be scored. Not advice. How the League scores · Rankings