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

MARKETS · Settled

Will above 100000 jobs be added in September 2026?

The answer: NO

The panel of 20 models averaged 62% YES; 5 of 20 were on the right side. The crowd price when the panel was sealed: 35%.

Forecasting closed 2 Oct 2026, 12:29 UTC · settled 2 Oct 2026, 13:45 UTC

THE BRIEF, WRITTEN AFTER THE OUTCOME

Panel favoured YES, crowd (35%) and five models favoured NO; outcome NO

The outcome was NO. The panel mean was 62% and the median 63%, both higher than the crowd Polymarket price of 35%. Five of 20 models called the NO outcome; the best-performing caller was Llama 4 Maverick + web search at 36% (near-crowd). The largest misses were high-YES forecasts such as DeepSeek R1 at 85%.

  • Outcome: NO, resolved 2026-10-02.
  • Crowd: Polymarket price frozen at 35%, it predicted NO.
  • Panel: 20 models, 5 called it correctly, mean 62%, median 63%.
  • Lowest panel YES probability 36%, highest 85%, disagreement span 49 points.
  • Best caller brier 0.13 (Llama 4 Maverick + web search); worst brier 0.723 (DeepSeek R1).

EVERY FORECAST, HIGHEST FIRST

  • DeepSeek R185%
  • Gemini 2.5 Flash78%
  • DeepSeek V378%
  • Perplexity Sonar Pro78%
  • Gemini 2.5 Pro76%
  • GPT-574%
  • Claude Sonnet 4.574%
  • Kimi K272%
  • Claude Opus 4.168%
  • Grok 4.563%
  • Perplexity Sonar63%
  • GPT-5 Mini62%
  • Mistral Large62%
  • GPT-5 + web search55%
  • Llama 4 Maverick55%
  • Qwen3 235B45%
  • Claude Sonnet 4.5 + web search38%
  • Gemini 2.5 Pro + web search38%
  • DeepSeek V3 + web search38%
  • Llama 4 Maverick + web search36%

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