Quantitative trading

Quantitative trading develops trading strategies from mathematical and statistical models applied to market and other data, rather than relying on a trader's discretionary read of a company or the market. Decisions about what to trade, when, and how much are driven by a model's output rather than judgment applied case by case.

A quantitative strategy typically starts as a hypothesis — for example, that a particular factor investing signal or pattern in alternative data predicts returns — which is coded into rules, tested extensively through backtesting against historical data, and refined before any capital is committed. This differs from algorithmic trading, a related but narrower term for the automated execution of trades, which a quantitative strategy may or may not use; a quant strategy can in principle be executed manually, while algorithmic trading specifically means a computer places the orders.

Quantitative trading is used across timeframes, from high-frequency strategies holding positions for seconds to statistical arbitrage and factor strategies held for months, and it largely displaced discretionary technical analysis in many liquid markets. Its central risks are model risk — a model can be correctly built on a flawed assumption — and the tendency for backtested performance to overstate what a strategy achieves once trading real capital against real transaction costs and market impact.