Our public research agenda is designed to explain how we think, not to disclose live trading strategies. StatMind publishes selectively, sharing research principles, methodology, and high-level findings while keeping live signals, model specifications, trade rules, data sources, and risk parameters private.
Our current areas of focus.
StatMind studies the full lifecycle of market research: from probability estimation and market-implied pricing to execution quality, monitoring, and risk-aware capital allocation.
Probability Modeling
Estimating fair value across uncertain, fast-moving market states with an emphasis on probability quality, calibration, and uncertainty measurement.
Market-Implied Pricing
Studying how public prices reflect information, liquidity, timing, and participant behavior.
Residual Edge
Understanding when disagreement between independent estimates and market-implied prices may be meaningful, noisy, or constrained by execution.
Execution Quality
Evaluating fill probability, queue position, slippage, liquidity, and adverse selection to understand the gap between theoretical and realized edge.
Model Monitoring
Tracking calibration, prediction distribution shifts, feature stability, regime changes, and post-deployment performance.
Risk-Aware Systems
Connecting signal quality to sizing, exposure, drawdown constraints, and strategy-level attribution.
Research discipline before deployment.
StatMind approaches markets through hypothesis generation, temporal validation, calibration review, failure analysis, and controlled deployment. The goal is to separate durable signal from noise before capital is put at risk.
Hypothesize
We begin with a testable market behavior, pricing question, or model failure.
Backtest
We evaluate ideas against historical data while accounting for timing, availability, leakage, and market context.
Validate
We use temporal validation, out-of-sample testing, and champion/challenger comparisons.
Calibrate
We evaluate whether model probabilities reflect reality, not just whether directional predictions appear accurate.
Monitor
We track live behavior across model quality, data health, market conditions, and execution outcomes.
Attribute
We decompose results to understand whether outcomes came from signal quality, market movement, execution, sizing, or noise.
Public notes are forthcoming.
StatMind will publish selected research notes, market structure observations, model writeups, and engineering essays over time. We will not publish live strategy details, sensitive model specifications, proprietary trade rules, or information that compromises active research.
Interested in the same problems?
We are open to serious conversations with researchers, engineers, traders, and market structure operators working on probability, prediction markets, execution systems, and risk.