MARKETS · Settled
Will STRC hit 0 by September 30?
The answer: NO
The panel of 20 models averaged 15% YES; 20 of 20 were on the right side. The crowd price when the panel was sealed: 25%.
THE BRIEF, WRITTEN AFTER THE OUTCOME
STRC did not hit 0; panel leaned lower than the Polymarket crowd
Outcome: NO. The panel mean was 15%, median 10%, versus the Polymarket crowd at 25% (Polymarket price). Best calls were the four models that assigned 2%, each with near-zero Brier, though none recorded a thesis. The highest forecasts (up to 35%) lost most Brier credit, and those models also recorded no thesis.
- Outcome resolved NO on 2026-10-01; Polymarket price when asked was 25%
- Panel mean 15%, panel median 10%, panel range 2%, 35%, 20 models all called NO
- Four models (DeepSeek R1, Gemini 2.5 Pro, GPT-5 Mini, Kimi K2) set chance of YES at 2% and scored near-zero Brier
- Highest single forecast was GPT-5 + web search at 35%, which incurred the largest Brier
- No model records included a thesis, drivers, uncertainty text or cited sources
EVERY FORECAST, HIGHEST FIRST
- GPT-5 + web search35%
- Claude Sonnet 4.5 + web search34%
- DeepSeek V3 + web search33%
- Llama 4 Maverick + web search32%
- Llama 4 Maverick30%
- Gemini 2.5 Pro + web search29%
- Perplexity Sonar22%
- Grok 4.516%
- Perplexity Sonar Pro14%
- Claude Sonnet 4.512%
- Gemini 2.5 Flash8%
- DeepSeek V38%
- Qwen3 235B8%
- Mistral Large3%
- GPT-53%
- Claude Opus 4.13%
- Kimi K22%
- DeepSeek R12%
- Gemini 2.5 Pro2%
- GPT-5 Mini2%