BestEx Research Debuts Predictive Cost Model for US Equities

Traditional models often struggle to isolate the impact of intraday timing, frequently conflating execution speed with order size. BestEx addresses this by utilizing a hybrid methodology that combines over one million institutional parent orders with transaction cost sensitivities derived from TAQ data. In testing, the model’s predictions remained within a fraction of a basis point of realized execution costs across diverse market conditions and liquidity buckets.

Founder and CEO Hitesh Mittal noted that robust predictive capabilities are essential for avoiding errors in position sizing. Inaccurate cost projections can lead to the rejection of viable trades or the under-sizing of attractive positions. By providing reliable pre-trade estimates, the system allows managers to refine their execution strategies before entering the market.

Clients can access the model through the Pulse AI assistant, a REST API, or the AMS One point-and-click interface. The suite includes four years of historical trading analytics, covering metrics such as volatility, spread, and depth. BestEx plans to expand its cost estimation coverage to global equities, with an AI-ready post-trade analysis capability expected in early 2027.