Whether you are an institutional portfolio manager, or a trader with quantitative systematic trading models, optimal order execution is critical.
Traditionally, execution quality is considered after the fact. Transaction Cost Analysis (TCA) is typically used to measure execution performance versus a benchmark such as the market price at time of order submission (“Arrival Price”) or the weighted average execution price in the market over the duration of the order (“VWAP”).
TCA is a very useful tool but it’s use is often motivated by the need to satisfy regulatory needs (e.g. MIFID II) or to compare relative broker execution performance.
Optimizing execution quality requires the following key elements:
- Ongoing measurement
- Market characteristics (e.g. volume, direction, volatility) continually change. Therefore, measurement of execution quality should be continuous and part of the workflow of order execution.
- Dynamic selection of execution algos
- Because markets change, execution results from a given algo will also change. Consequently, decisions on algo selection should be dynamic, based on market conditions.
- Market data is essential
- Dynamically selecting execution algos and measuring performance requires level 1 data (best bid/offer, trades) at the minimum. Level 2 data (market depth) improves accuracy and effectiveness.
Order Execution Products
ExchangeSimulator operates on a knowledge basis of orders, executions and market data from each firm. We call this technology and knowledge base “AlgoGrid”.
ExchangeSimulator specifically addresses the needs of algo developers to test the operation of algos against an exchange: simulating a real-time trading environment and back-testing against historical market data. Architecturally, it resides within ExecutionServer.
ExchangeSimulator includes market data adapters for Level 1, Level 2 and Level 3. Market data modified by simulated orders is published via FIX or Deltix multi-cast API
The simulator maintains an order book for each instrument. Orders sent for simulated execution by the execution strategies are filled according to the volume available in the order book (at the appropriate price levels in the case of a limit order) at sub-millisecond intervals from the time the order is received. The removal of liquidity from the order book by the simulated execution is reflected in the state of the order book.
Individual execution strategies/algos control the sending of orders to the simulation engine. The state of the order book may determine if and when the algo sends an order (e.g. Volume participation strategies). Latencies can be dialed up and down to simulate the actual production environment.
All Deltix products support Equities, Futures, Bonds, ETF, FX, and Synthetics.
Deltix ExecutionServer is a platform for executing manual orders or systematic orders generated by automated trading strategies. It provides the following services with “wire-to-wire” single-digit μs latency:
- Execution Algos
- Trading risk management
- Exchange Simulator
- Market Making
- Transaction Recording / Audit
- Ultra-low latency High Availability.