Platform

Our work runs on Longleaf, a trading platform we designed and built for strategy research, backtesting, and live execution.

Strategy research

Longleaf is written in OCaml. Strategies are expressed in a typed strategy language built on it, so errors in strategy logic are caught before a strategy ever runs. Backtests cover single periods or batteries of historical periods using minute-level market data, and strategy parameters can be tuned with numerical optimization.

Fast backtesting

Backtests run at hundreds of thousands of bars per second on a single core. A year of minute-level data across the S&P 500 completes in under two minutes, data loading included, several times faster than comparable engines. Ideas can be tested, refined, and tested again in minutes.

Paper and live trading

The same engine runs backtests, paper trading, and live trading, so a strategy behaves the same way in testing and in production. Equities, ETFs, and futures all trade through one engine, and brokerages connect through a common interface.

Monitoring and control

A web interface shows running strategies as they trade. Statistics can be saved at any point, and a run can be stopped with positions left in place or fully liquidated.