Algorithmic execution is the discipline of getting a decided trade into the market at the lowest total cost — the automated machinery that takes a "parent" order (buy 500,000 shares of X) and works it into the market as a stream of small "child" orders, deciding how much to trade, when, and where. It is deliberately downstream of the alpha decision: by the time an execution algo runs, the choice to buy or sell is already made, so its objective is not to predict direction but to minimize the cost of converting that decision into a filled position. Its governing tension is the "trader's dilemma" — trade fast and you move the price against yourself (market impact); trade slow and you risk the price drifting away while you wait (timing/opportunity cost). Every algorithm in this branch is, at bottom, a different answer to that single trade-off (QuestDB; arXiv 1206.5324; CFA/AnalystPrep trade-execution notes).

What this section covers

This is the execution layer of quantitative trading — distinct from, and one level below, signal generation. It assumes a position has already been chosen and asks the narrow but economically large question: what is the cheapest way to actually trade it in a market that is fragmented across venues, populated by faster participants, and where your own demand for liquidity moves the price? The two structural cost drivers it manages are:

• Market impact — the price moves against you the more aggressively you demand liquidity (temporary impact that decays, plus permanent impact that does not).
• Timing risk — while you trade patiently, the price can drift away on its own, and shares you never fill forfeit the alpha you were chasing.

These two costs pull in opposite directions and cannot be simultaneously minimized; the art of execution is choosing where on that frontier to sit, given urgency and risk tolerance.

Why it matters — and when it doesn't

Execution cost scales with order size relative to a name's liquidity. For an institution accumulating multiples of a stock's average daily volume, implicit costs (impact + delay + missed trades) routinely dwarf commissions — a frequently cited Plexus Group dataset reproduced in the CFA curriculum found delay alone at 84 of 153 total basis points of shortfall on one institutional equity sample (one region, one period; not a universal figure). At that scale, execution quality is a material, measurable component of net return.

For a retail-scale order in a liquid stock — a few hundred shares of a high-volume name — the entire framework largely collapses to "pay the spread." The order is small relative to displayed liquidity, market impact is negligible, and the broker's smart router handles venue selection invisibly. The honest scope boundary: algorithmic execution is a large-order / illiquid-name discipline. Forcing its machinery onto small, liquid trades adds complexity without saving meaningful cost.

Map of the sub-topics

This section's children cover the three pillars — the scheduling algorithms, the venue layer beneath them, and the benchmark that scores the whole thing:

• TWAP / VWAP Execution (001) — the foundational schedule-based algorithms. TWAP slices an order into equal pieces at uniform time intervals; VWAP releases slices in proportion to forecast intraday volume so fills track the volume-weighted average price. Both are passive, benchmark-tracking, and direction-agnostic — they answer the how much / when question while largely ignoring the trader's dilemma (they accept timing risk to minimize signaling). Best for non-urgent orders. The child also separates the genuine debate (VWAP as a during-trade benchmark is gameable and arguably inferior to a pre-trade benchmark) from the uncontested mechanic (slicing reduces impact versus block trading).

• Smart Order Routing (002) — the sub-second layer beneath the scheduling algos that answers where each child slice executes. In a Reg-NMS market fragmented across roughly a dozen-plus lit exchanges, dozens of dark pools, and wholesale market makers, SOR scans venues and routes to honor the Order Protection Rule (no trade-throughs) while sourcing hidden liquidity. Its central, well-documented controversy is the conflict of interest from maker-taker rebates and payment for order flow: the routing logic optimal for the broker is not always optimal for the client, which is why SEC Rules 605/606 exist.

• Implementation Shortfall (003) — both a benchmark and an algorithm family. As Perold's (1988) TCA metric, it is the most complete cost measure: the gap between the price at the moment of decision and the price actually realized, including shares never filled, decomposed into explicit cost, execution slippage, delay, and missed-trade opportunity cost. As an algorithm ("arrival price"), the IS family — canonically modeled by Almgren–Chriss (2000) — explicitly optimizes the trader's dilemma, front-loading to trade off expected impact against timing-risk variance. This is the child that confronts head-on what TWAP/VWAP ignore.

The natural reading order is IS first (it defines the cost the whole branch is trying to minimize), then TWAP/VWAP (the simple schedule-based answers), then SOR (the venue substrate every algo runs on top of).

Strengths & limitations of the domain

Strengths. Algorithmic execution reliably reduces spread-and-impact cost versus naive block trading, camouflages large orders within normal flow, produces an auditable best-execution trail, and — through IS-style TCA — makes hidden costs (delay, unfilled alpha) visible that commission-only accounting conceals entirely.

Limitations. Every layer adds latency and complexity, and at execution speed the algos compete with HFT, exposing slower routers to adverse selection and information leakage (multi-venue pinging can telegraph a large order and increase impact). The cost models themselves rest on assumptions — impact functions are nonlinear, transient, and regime-dependent — so an "optimal" trajectory is optimal only under its model. And the benchmarks can mislead: the single most common error across the whole branch is reading one order's result as a verdict on execution skill, when a single trade's cost is mostly volatility noise. TCA is meaningful only in aggregate, over hundreds of orders.

Sources

• QuestDB glossary — Algorithmic Execution Strategies (definition; impact-vs-timing trade-off).
• AnalystPrep / CFA Level III — Trade Execution notes (trader's dilemma, scheduling vs arrival-price algos).
• arXiv 1206.5324, Effective Trade Execution (trade-scheduling dilemma: fast = market risk vs slow = slippage).
• Perold, A. (1988), The Implementation Shortfall: Paper Versus Reality (the cost the branch minimizes).
• Almgren, R. & Chriss, N. (2000), Optimal Execution of Portfolio Transactions (efficient frontier of execution trajectories).
• Plexus Group (2004) via the CFA Institute curriculum — delay-cost figure (one institutional sample; illustrative, not universal).
• Child docs in this section: TWAP/VWAP Execution (001), Smart Order Routing (002), Implementation Shortfall (003) — fully sourced individually.

Confidence: medium. This is a section-overview altitude doc; the precise formulas, base rates, and contested-evidence detail live in the three child docs and are sourced there. The one numeric figure cited here (84/153 bps) is qualified as a single-sample illustration, not a market average.