An execution algorithm helps carry out a trade that has already been chosen. It decides how to distribute orders over prices or time. It does not establish that the original position is worth taking.
The central trade-off is simple: a large immediate order may move the price against you, while a slower order leaves more exposure to what happens before completion.
The cost of immediate completion
Suppose you need to buy 600 units. Sellers offer 200 at $50, 200 at $50.10 and 200 at $50.20. Buying them immediately costs $30,060, an average of $50.10.
Relative to an initial $50 decision benchmark, the execution cost is $60, or 0.20%—20 basis points—of the $30,000 benchmark notional. The market may replenish after the first slice, but it may also rise before the rest can execute.
If you instead buy only 200 at $50 and leave 400 unfilled, the visible fill looks better. If the deadline arrives with price at $50.50, buying the remainder then costs another $200 relative to the original benchmark. If the remainder is never bought, that $200 is an opportunity-cost comparison, not a cash loss recorded by the account.
Time and volume schedules
A time-weighted average price schedule, or TWAP, spreads an order across a chosen interval. It is simple to specify, but equal time slices can encounter very different available liquidity.
A volume-weighted schedule, or VWAP, aims to trade more when market volume is higher. It depends on the quality of the volume estimate or participation data. The market can trade heavily while moving sharply against the order, so matching a volume pattern does not ensure an attractive absolute price.
CME’s transaction-cost analysis explains why the schedule, urgency and evaluation benchmark need to match the objective. Beating the session’s VWAP can coexist with paying more than the price when the decision was made.
A chase order follows quotes, not your valuation
A chase order repeatedly adjusts a limit price toward a moving quote. It can reduce the manual work of staying near the market, but it introduces a critical question: how far may the order follow?
Insilico’s chase documentation distinguishes price-constrained behavior from an until-fill mode that removes the adverse-price cap. That difference changes the economic instruction. “Stay competitive while the price remains acceptable” is not equivalent to “keep following until the quantity fills.”
A passive order also faces selection: it may fill more readily just as price is moving against its owner. A higher fill rate is not automatically better execution if the fills occur in systematically worse conditions.
The parent order and its children
The parent instruction describes the total trade. Child orders are the individual orders it sends. Track filled quantity, active quantity and the quantity still scheduled. A partially completed algorithm can have several kinds of commitment at once.
Set the maximum quantity, worst acceptable price where applicable, deadline and behavior when liquidity is insufficient. Decide whether the deadline cancels the remainder or permits more aggressive completion. Those are different trades, not interchangeable convenience settings.
After stopping an algorithm, confirm which existing child orders remain active. Stopping new submissions need not cancel working orders, and cancelling orders does not close a position already filled.
Evaluate the full decision against a stated benchmark: fills, explicit costs and any uncompleted quantity. An algorithm that reports cheap fills while omitting the uncompleted quantity can look successful precisely because it avoided the expensive part of the trade.