Optimal Execution

Optimal execution is the problem of minimizing the cost of trading a large order over time, subject to the constraint that the trading itself must not reveal so much information that the market moves against the trader. It is not a problem of finding the best price at a single moment; it is a problem of managing the information leakage that every trade produces across the entire execution horizon.

The naive approach — breaking a large order into equal-sized chunks and trading them at fixed intervals — is demonstrably suboptimal. It ignores the temporal structure of market microstructure noise, the strategic behavior of other algorithms, and the possibility that the market's liquidity profile changes throughout the day. Optimal execution algorithms instead model the tradeoff between market impact and timing risk, treating the order book as a dynamic system rather than a static queue.

The field sits at the intersection of stochastic control, game theory, and market design — and it is notable for how often the theoretically optimal strategy differs from what practitioners actually do, not because practitioners are irrational, but because the theoretical models assume a stationary environment that real markets do not provide.