Free lesson · Trading algorithms

TWAP, VWAP and percentage-of-volume execution

Open interactive lesson Practice calculations Explore labs

Start with the idea

Execution algorithms schedule a parent order. They usually aim to balance urgency, benchmark tracking and liquidity consumption; they do not by themselves decide whether the investment should be held.

Symbols, units & horizon

Q: parent quantity
N: time buckets
q_i: scheduled quantity in bucket i
V̂_i: forecast volume for VWAP scheduling
V_external: observed external volume in this teaching POV convention
ρ: target fraction of external volume
Actual fills: may differ from scheduled quantities
Final bucket: handle rounding and remaining quantity

When and why to use this

Use scheduling algorithms to implement a parent order consistently and compare benchmark tracking, cost and completion risk.

TWAP spreads quantity across time buckets, making a simple schedule without a volume forecast. VWAP-oriented scheduling follows an estimated intraday volume profile. Percentage-of-volume execution adapts to observed market volume, subject to completion and participation constraints.

The VWAP benchmark is a realised volume-weighted price; a prospective VWAP schedule must use volume estimates available then. Using the day’s final volume curve to schedule earlier trades is hindsight. POV’s denominator convention matters: some participation definitions include your own prints, while others use external volume.

Choose based on urgency, liquidity profile, information decay and benchmark. A passive schedule can miss completion; an urgent schedule pays spread and impact. Commercial implementations include venue, order-type and risk logic beyond the teaching formulas, so actual behaviour must be read from current broker specifications.

q_{i}^{T W A P} = \frac{Q}{N}, q_{i}^{V W A P} = Q \frac{{\hat{V}}_{i}}{\sum_{j} {\hat{V}}_{j}}, q_{i}^{P O V} = ρ V_{i}^{e x t e r n a l}

Model assumptions, derivation and arithmetic

TWAP, VWAP and percentage-of-volume execution

TWAP allocates equal fractions of Q across N intervals.
VWAP normalises the forecast volume profile to fractions summing to one, then multiplies by Q.
POV multiplies observed external volume by the desired fraction and caps at the remaining order. If participation includes own volume, solve q/(V_external+q)=ρ, giving q=ρV_external/(1−ρ) instead.

Work it by hand

For Q=1000 and forecast volumes [1,2,1], VWAP schedule is [250,500,250], while TWAP is approximately [333.33,333.33,333.33]. A 10% external-volume rule trades 200 against 2000 external units, capped by remaining quantity.

Apply it in a strategy

State parent quantity, deadline, benchmark, participation denominator and allowed order types.
Forecast volume from past comparable sessions or use a causal realised-volume rule.
Backtest scheduled versus filled quantities, slippage and unfilled terminal obligations under the same market conditions.

Research deliverable

Create a schedule and fill ledger for TWAP, forecast-VWAP and POV, including rounding, completion and cost assumptions.

Python implementation

Self-contained teaching example. Python 3.10+; dependencies and input conventions are shown in the code and notation. Run in your own Python environment.

def schedules(quantity,volume_forecast):
    if quantity<0 or not volume_forecast or any(v<0 for v in volume_forecast) or sum(volume_forecast)<=0: raise ValueError("Valid schedule inputs required")
    return [quantity/len(volume_forecast)]*len(volume_forecast),[quantity*v/sum(volume_forecast) for v in volume_forecast]

def pov_external(volume,participation,remaining):
    if min(volume,participation,remaining)<0 or participation>1: raise ValueError("Invalid participation inputs")
    return min(remaining,participation*volume)

print(schedules(1000,[1,2,1]),pov_external(2000,.1,1000))

Continue learning

Trading Algorithms: A Practical Selection Guide — all lessons

Quantitative finance and development glossary · Python resources and libraries · Research sources and limitations