ML volume forecasting turns trade execution into alpha, Goyenko/Kelly/Moskowitz study
iblanco_finance · x · 2026-11-13
iblancofinance breaks down the "Trading Volume Alpha" research: predicting which stocks will trade heavily is worth about as much to a portfolio as predicting returns, since trading costs hinge on volume around execution — and volume is highly predictable.
Ruslan Goyenko (McGill), Bryan Kelly and Tobias Moskowitz (Yale SOM/AQR/NBER) et al. train ML models on past volume, price and liquidity data to forecast volume stock by stock, then plug forecasts into a portfolio rule trading off tracking error against net-of-cost performance. The author's read: execution isn't an afterthought — it's part of the alpha. SHAP analysis in the figure shows technical predictors dominate, followed by calendar and earnings features.
More from Venture
- Mantle Launches AI-Native Maintenance Service for Robots and Industrial Machines With 250+ Technicians — ycombinator · 2026-11-13
- "RIP Two Sigma": viral LinkedIn post claims a $200 Claude Code plan rebuilt a quant desk — PtrPomorski · 2026-11-13
- Toronto-built Ribbon sold for life-changing money, founder urges local builders to aim global — AJChadha · 2026-11-13
- Voice AI firm Retell hits $100M ARR, up 10x in 17 months, profitable on $4.6M raised — s_mohinii · 2026-11-13
- Biotech's biggest funding vacuum: $10-15M meaningful Seed rounds nobody provides — PatrickKidger · 2026-11-13
- Indie product hits 134 HQ users in 4 months as growth curve goes vertical — jacob_posel · 2026-11-13