This article deconstructs the core pillars of that framework, explaining how DMA provides the "highway" and algorithms provide the "driver" for modern electronic trading. Before algorithms could trade, they needed a direct line. Traditional brokerage involved calling a human broker who would execute your order, often via the firm’s proprietary desk. DMA flips this model.
| Strategy Type | Primary Goal | Typical User | Example | | :--- | :--- | :--- | :--- | | | Track historical or intraday volume profile | Asset managers, pension funds | Execute 5% of volume every 10 minutes | | Implementation Shortfall | Minimize slippage vs. arrival price | Active long/short equity funds | Aggressive at start, passive thereafter | | Liquidity Seeking (Sniffer) | Find dark pool or hidden liquidity | High-turnover quant funds | Sweep odd lots, ping dark venues | | Market Making | Earn bid-ask spread; provide liquidity | Prop trading firms, HFT | Post limit orders, cancel on adverse move | The Critical Interface: Risk Management in DMA The most overlooked chapter in Johnson’s framework is risk . DMA removes the broker’s manual checks, placing immense responsibility on the buy-side firm. An algorithmic error without DMA might cause a slow loss; with DMA, it can cause a flash crash in your own account in seconds. algorithmic trading and dma barry johnson pdf
Barry Johnson categorizes algorithmic strategies into two high-level families: (seeking to track a benchmark like arrival price) and opportunistic (seeking to capture liquidity or arbitrage). This article deconstructs the core pillars of that