Quantitative Modeling • Market Microstructure • NFL Forecasting
Probabilistic models for competitive markets.
Vertical Insights builds custom forecasting models for complex markets, with emphasis on edge quantification, latent state tracking, and model–market differentials. Current work includes variational Bayesian models of pre-game and in-game NFL point spreads that decompose team performance into static strengths and time-varying factors.
About
Vertical Insights is led by Jacek Dmochowski, an Associate Professor of Biomedical Engineering (CUNY) and VP of Machine Learning at a neurotechnology startup (Optios). His research spans probabilistic modeling, signal processing, and applied machine learning for decision-making under uncertainty.
Peer-reviewed research
For a formal treatment of optimal wagering under distributional uncertainty, see “A statistical theory of optimal decision-making in sports betting” , PLOS ONE 18(6): e0287601, 2023.
Example Model Artifact
Latent team deviations over time
Posterior deviations (median) from the market spread for three franchises (DAL, KC, SF) over the 2025 season. Deviations are inferred from a variational Bayesian model incorporating fixed strengths and autoregressive latent dynamics.
Collaborations & Conversations
If you’re exploring collaborations, proprietary modeling work, or want to discuss approaches to market forecasting in more depth, feel free to reach out.