Skip to content

Level III Core · Trading Costs and Electronic Markets

Algorithmic and High-Frequency Trading for CFA Level III

Updated 8 October 2026 · Fact-checked

Algorithmic trading uses computer rules to split and place orders. Execution algorithms such as TWAP, VWAP and arrival price (implementation shortfall) schedule child orders against time, volume or the decision price. High-frequency trading is very fast, short-horizon trading. To solve questions, match the algorithm to the client's urgency, benchmark and risk.

Understand Algorithmic and High-Frequency Trading

An algorithm is a set of rules a computer follows to trade. Instead of sending one large order, the system breaks it into many small child orders. It sends them to different venues over time. The aim is to cut market impact and explicit costs while keeping control of risk.

There are two broad uses. Execution algorithms carry out a trade that a portfolio manager has already decided. Alpha-seeking or proprietary algorithms generate their own trade ideas, such as statistical arbitrage or market making. Level III focuses mainly on execution algorithms, because they link to your trade strategy and cost analysis.

The main execution algorithms differ in the benchmark they chase. TWAP (time-weighted average price) slices the order evenly across time. VWAP (volume-weighted average price) slices it in line with the expected intraday volume pattern, so you trade more when the market is busier. Arrival price (implementation shortfall) algorithms try to stay close to the price when the order arrived. They trade faster at the start to cut the risk of price drift, then slow down to limit impact. Percentage of volume (POV) algorithms trade a fixed share of market volume as it occurs. Liquidity-seeking algorithms look for hidden or dark liquidity and trade when size is available.

The trade-off is always market impact versus timing (opportunity) risk. Trading fast raises impact but lowers exposure to price drift. Trading slowly lowers impact but raises the risk that the price moves against you. A client who is urgent or has information that will decay should trade faster. A client with no view and low urgency can trade passively.

High-frequency trading (HFT) uses very fast, automated systems, often with co-located servers, to trade at speeds measured in fractions of a second. Typical strategies include market making, arbitrage across venues, and short-term statistical strategies. HFT can narrow spreads and add displayed liquidity in normal conditions. Critics argue that liquidity can vanish in stress, that some strategies are predatory toward slower orders, and that fast feedback loops can raise the risk of sudden price dislocations. Both sides are examined, so be ready to give a balanced view.

Key rules to remember

TWAP slice size
Slice per interval = Total order size ÷ Number of equal time intervals
TWAP ignores volume patterns. Use it when volume is hard to predict or you want a simple, even schedule.
VWAP
VWAP = Σ(price × volume) ÷ Σ(volume)
A VWAP algorithm targets this benchmark by trading in proportion to expected volume in each interval.
VWAP slice size
Slice in interval = Total order × (Expected volume in interval ÷ Expected total volume)
Uses the forecast volume profile, so accuracy depends on the forecast.
Percentage of volume
Order size traded = Participation rate × Market volume
Finishing time depends on actual volume, so completion is not guaranteed.
Implementation shortfall (buy)
Shortfall per share = Execution price − Decision (arrival) price
For a sell, reverse the sign. Include delay, impact, fees and unexecuted portions when the question asks for total shortfall.

How to solve Algorithmic and High-Frequency Trading questions

Use this method for any question on execution algorithms or HFT. Tie your answer to the client and the order.

  1. 1Read the command word and the scenario. Identify the order: size relative to normal volume, urgency, and whether the client has a view on price.
  2. 2Identify the benchmark the client or trader is judged against: arrival price, VWAP, close, or none.
  3. 3Decide the main risk. If price drift or information decay matters most, favour faster trading and arrival price. If impact matters most, favour slower, passive trading.
  4. 4Match the algorithm: TWAP for even timing, VWAP for volume-matched trading, POV for following live volume, arrival price for urgency near the decision price, liquidity-seeking for large or illiquid orders.
  5. 5If a calculation is asked, write the formula, substitute the numbers, and show each slice or price difference before the final number.
  6. 6For HFT questions, give both the liquidity benefit and the risk, then link to the scenario.
  7. 7State your recommendation in one sentence with one reason that points to the client's objective or constraint.

Quickest way: Urgency and benchmark shortcut

When to use it: Use this on item-set questions that ask which algorithm or strategy fits a described order.

  1. Urgent or information-driven order: pick arrival price or a faster, more aggressive algorithm.
  2. Benchmark is VWAP, or order is routine and not urgent: pick VWAP.
  3. Simple even schedule or unknown volume pattern: pick TWAP.
  4. Large order in illiquid stock: pick liquidity-seeking or low participation POV, and accept a longer time.
  5. Eliminate options that ignore the stated urgency or benchmark.

Common mistakes in Algorithmic and High-Frequency Trading

  • Treating TWAP and VWAP as the same thing.

    Both spread an order over the day, so they look alike.

    Fix: TWAP slices by equal time. VWAP slices by expected volume. Say which input each uses.

  • Recommending VWAP for an urgent, information-driven order.

    VWAP is the most familiar algorithm, so students pick it by default.

    Fix: VWAP spreads trades across the day and leaves the order exposed to drift. Choose arrival price or a faster approach when urgency is high.

  • Saying HFT is simply good or simply bad for liquidity.

    Students memorise one side of the debate.

    Fix: Give both: tighter spreads and more displayed depth in normal markets, but liquidity that can withdraw in stress and risks from speed-based strategies.

  • Forgetting the impact versus timing risk trade-off.

    Students focus on the algorithm names, not the reason for choosing one.

    Fix: State that faster trading raises impact and lowers timing risk, and slower trading does the reverse. Link it to the client.

  • Assuming a POV algorithm always completes the order.

    The participation rate sounds like a guaranteed schedule.

    Fix: POV trades a share of actual volume. If volume is low, the order may not finish in time.

  • Getting the sign wrong on implementation shortfall.

    Students subtract in the same order for buys and sells.

    Fix: For buys, shortfall is execution price minus decision price. For sells it is decision price minus execution price. A positive number is a cost.

Worked examples

Example 1

A trader must buy 60,000 shares over a 6-hour session using a VWAP algorithm. Expected volume shares of the session are: hours 1-2, 40%; hours 3-4, 20%; hours 5-6, 40%. How many shares should the algorithm buy in each two-hour block? How would a TWAP schedule differ?

Show the solution
  1. VWAP slice = Total order × expected volume share.
  2. Hours 1-2: 60,000 × 0.40 = 24,000 shares.
  3. Hours 3-4: 60,000 × 0.20 = 12,000 shares.
  4. Hours 5-6: 60,000 × 0.40 = 24,000 shares.
  5. Check: 24,000 + 12,000 + 24,000 = 60,000.
  6. TWAP splits by equal time. Three equal two-hour blocks give 60,000 ÷ 3 = 20,000 shares per block.

Answer: VWAP: 24,000, 12,000 and 24,000 shares. TWAP: 20,000 shares in each block.

Example 2

A portfolio manager decides to buy a stock when its price is 250.00. The order is executed in full at an average price of 250.60. Fees total 0.05 per share. What is the implementation shortfall per share for 10,000 shares, and which algorithm type is aimed at reducing it?

Show the solution
  1. For a buy, price shortfall per share = 250.60 − 250.00 = 0.60.
  2. Add fees per share: 0.60 + 0.05 = 0.65.
  3. Total shortfall = 0.65 × 10,000 = 6,500 in the stock's currency.
  4. The algorithm type that targets the decision or arrival price is the arrival price (implementation shortfall) algorithm.

Answer: Implementation shortfall is 0.65 per share, or 6,500 in total. An arrival price algorithm is designed to reduce it.

Exam tips

  • For recommend or justify questions, give the algorithm and one reason tied to urgency, size or benchmark. Extra reasons rarely add points.
  • In calculations, write the formula and each slice. If the command word is calculate, the correct number alone earns credit, but showing steps protects you if you slip.
  • Know the HFT debate in both directions. Item sets often test a single claim, so read each option for overstated words like always or never.
  • Match the algorithm to the stated benchmark. If the scenario names VWAP as the benchmark, a VWAP algorithm is usually the fit.
  • Answer only the number of points asked for, in the order given. Extra responses are not evaluated.

Algorithmic and High-Frequency Trading in other exams

The same ground in other exams, if you are preparing for more than one or want another angle on it.

Algorithmic and High-Frequency Trading: frequently asked questions

What is the difference between TWAP and VWAP?

TWAP splits an order evenly across time. VWAP splits it in proportion to expected volume, so it trades more in busy periods. VWAP aims to match the volume-weighted average price. TWAP is simpler and useful when volume is hard to forecast.

What is an arrival price algorithm?

It aims to keep the average execution price close to the price when the order arrived. It usually trades more at the start to reduce drift risk and then slows to limit market impact. It suits urgent or information-driven orders.

How does high-frequency trading affect market liquidity?

In normal conditions HFT market making can narrow spreads and add displayed depth. In stressed markets some HFT firms may pull quotes, so liquidity can disappear quickly. Give both sides in exam answers.

What are the main types of execution algorithms?

The common types are TWAP, VWAP, percentage of volume, arrival price or implementation shortfall, and liquidity-seeking algorithms. Each balances market impact against timing risk in a different way.