Introduction to Algorithmic Execution

Introduction to Algorithmic Execution - Part 1: Introduction

Published by: OrderX

•

Updated On: Jul 7, 2026

•

Introduction to Algorithmic Execution

A brief history of execution algorithms, what they actually do, and how the 14 parts of this series fit together.

Every trade you decide to put on still has to be turned into actual order fills - and how you do that determines a surprisingly large share of your PnL. This series is a practical course in algorithmic execution: the systems that take a large order and work it into the market intelligently. In this first part, we look at where execution algorithms came from, what they actually are, and how the rest of this series fits together.

From Floor Traders to Code

Not long ago, execution was a human craft. Through the late 1990s a large fraction of equity volume was still negotiated by people on exchange floors. Even electronically routed orders were mostly sized, timed, and priced by a human. Genuine machine-managed execution was a niche capability inside a handful of quantitative firms.

Electronification changed that within a decade. Once order books became fully electronic and market data machine-readable, it was natural to encode a trader’s pacing and pricing decisions into software. By the mid-2000s institutional desks had adopted execution algorithms at scale, and today they are the default in pretty much every market. “Algorithmic execution” has simply become execution.

Two forces keep pushing this adoption forward. First, economies of scale: a good algorithm encodes the best version of a decision once and applies it to every order thereafter. Second, measurement: because algorithms are systematic, their performance can be analyzed and improved in a way that discretionary trading cannot. We cover this in detail in Part 12: Benchmarks & TCA.

What Exactly Is an Execution Algorithm?

An execution algorithm is a program that converts a parent order (e.g. “buy 250,000 units by 4pm”) into smaller child orders, each with a size, price, venue, and timestamp. Execution algorithms can be categorized by:

By Scope

Single-order algorithms treat every parent order independently. Give them ten orders and they will run ten separate, unaware executions.
List (portfolio) algorithms execute a basket jointly, coordinating pace and risk across all traded instruments at once. We dedicate Part 8 to these.

By Function

Smart order routers answer a narrow question: given a child order right now, which venue (or combination of venues) should receive it? They are covered in depth in Part 9.
Scheduling algorithms answer the broader question of how to spread a parent order over time - and typically delegate venue choice to a router underneath.

By Pacing Style

Schedule-driven: follow a plan fixed at the start - TWAP, VWAP, and the classic implementation-shortfall trajectory in Part 10.
Dynamic: recompute the pace continuously from live conditions, like the Percent of Volume family.
Opportunistic: stay dormant and act only when conditions are unusually favorable, like the Hide and Take style. Hybrids add a fallback pace so the order can never stall entirely.

Building Blocks of every Execution Algorithm

Whatever the high-level strategy, almost every algorithm is assembled from the same reusable components, each configured differently per strategy:

Scheduler - decides how much should be done by when, and how far ahead or behind plan the order may drift.
Pricing logic - decides how aggressively each child order is priced, from resting passively at the bid to crossing the spread. Explored in Part 11.
Display logic - decides how much size is visible, hidden, or held in reserve.
Router - picks the destination venue(s) for each child order.

A simple mental model: a TWAP engine slices the parent evenly through time; for each slice the pricing logic picks a level, the display logic picks a visible quantity, and the router dispatches it. The loop then watches the market and re-sizes, re-prices, or re-routes until the order completes or is cancelled.