Stanford Paper Uses CNN on Raw Prices to Predict S&P 500, Challenging Factor Tradition
mdancho84 · x · 2026-10-12
A Stanford paper by Rahul Gupta challenges quant finance's long-held consensus that raw prices are too noisy and require hand-crafted indicators. It feeds multivariate raw data (including split/dividend events) into a CNN as image-like matrices to predict S&P 500 movements, reporting promising results across single stocks, sectors, and portfolios.
More from Venture
- The Circular AI Economy: Why OpenAI's Investor Chain May Be Too Big to Fail — CartoonistVisual2818 · 2026-10-12
- VCs spent two years staffing media teams — a sign attention, not capital, is scarce — damianplayer · 2026-10-12
- Anthropic subscribers reportedly get 4X more compute per dollar than API buyers as margins hit 88% — rohanpaul_ai · 2026-10-12
- VC argues angel investors need 50-100 company portfolios to hit a big winner — dunkhippo33 · 2026-10-12
- Man jailed 18 months for using bots to stream his AI-generated songs, earning over $8M — skdh · 2026-10-12
- Manus Hit $100M ARR in 8 Months: Agents Will Reward Product Quality Over Attention — parker_lyman · 2026-10-12