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

MARKETS · Settled

Will above 20000 jobs be added in September 2026?

The answer: YES

The panel of 20 models averaged 79% YES; 20 of 20 were on the right side. The crowd price when the panel was sealed: 82%.

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

THE BRIEF, WRITTEN AFTER THE OUTCOME

Panel correctly forecast YES, mean 79% vs crowd Polymarket 82%, wide model spread

The outcome was YES. The crowd was the Polymarket price at 82% and the panel mean was 79%, median 82%, with all 20 models calling YES. DeepSeek R1 made the strongest call at 96% and scored the lowest Brier, its record contains no thesis or drivers. The lowest forecast was Llama 4 Maverick at 58%, also with no recorded thesis or drivers.

  • Outcome: YES, resolved from Kalshi market KXPAYROLLS-26SEP-T20000.
  • Crowd: Polymarket price frozen at 82%, it called the outcome correctly.
  • Panel: 20 models, mean chance of YES 79%, median chance of YES 82%, all 20 called YES.
  • Range: lowest 58% (Llama 4 Maverick), highest 96% (DeepSeek R1).
  • Best Brier: DeepSeek R1 brier 0.002; worst Brier: Llama 4 Maverick brier 0.176.

EVERY FORECAST, HIGHEST FIRST

  • DeepSeek R196%
  • GPT-595%
  • Kimi K292%
  • GPT-5 + web search91%
  • GPT-5 Mini88%
  • Claude Opus 4.188%
  • Perplexity Sonar Pro88%
  • Claude Sonnet 4.585%
  • DeepSeek V385%
  • DeepSeek V3 + web search82%
  • Grok 4.582%
  • Llama 4 Maverick + web search77%
  • Gemini 2.5 Pro75%
  • Gemini 2.5 Pro + web search71%
  • Gemini 2.5 Flash70%
  • Mistral Large68%
  • Claude Sonnet 4.5 + web search68%
  • Qwen3 235B65%
  • Perplexity Sonar63%
  • Llama 4 Maverick58%

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