How Deep To Place A Limit Order
The trade-off an execution algorithm faces when resting a limit order: closer to the touch fills faster but risks adverse selection, deeper in the book waits longer but risks missing the trade entirely.
Prerequisites: Modelling Fill Probability for a Resting Order, Depth At Touch And The Shape Of The Book
Every time an execution algorithm decides to rest a passive order instead of crossing the spread, it faces a second, subtler choice: exactly where in the book to place it. Sit right at the best bid or offer and the order is near the front of the queue, likely to fill soon — but that's also where informed traders are most likely to trade against it. Sit several ticks back and the order is safer from that risk, but may never fill at all if the price moves away first.
Depth as a dial between speed and safety
Placement depth describes how far from the current best price a limit order rests — at the touch, or some number of ticks behind it. An order at the touch has the shortest queue to wait through and the highest chance of filling, but it's also first in line to trade against anyone with better information about where the price is headed — a cost known as adverse selection. An order placed deeper avoids that immediate exposure and tends to capture a better price if filled, but at a lower probability of filling at all, since the price has to move toward the order rather than the order simply waiting its turn.
The right depth depends on urgency (see urgency and the risk-aversion parameter) and the shape of the order book — in a thin book with a queue that clears quickly, resting a tick back can still fill promptly; in a deep, slow-moving book, the same depth might mean a long wait.
Worked example: comparing two depths on the same order
A stock is quoted $50.00 bid / $50.02 offer, and a trader wants to buy. Placing a limit order at $50.00 (the touch) puts it in the queue behind other resting orders; historical data suggests a touch order fills within two minutes about 70% of the time, but roughly a third of those fills happen just before the price ticks down further — the trader was often filled right as the market moved against the position. Placing the order one tick back at $49.99 drops the fill probability to about 35%, since the price now has to move down a tick to reach it — but the fills that do happen show less immediate adverse drift, because the price has already shown willingness to move toward the order. A trader with real urgency accepts the touch order's worse average fill quality for the higher chance of getting done; a patient trader accepts the lower fill probability at $49.99 for the better expected outcome conditional on filling.
What this means in practice
Placement depth is rarely set once and left alone — algorithms move an order deeper when they want to be more patient and closer to the touch as a deadline approaches, converting the passive-versus-aggressive trade-off into a continuous dial. It also depends on tick size: in a large-tick stock where the spread is nearly always one tick, each depth level matters a great deal; in a small-tick stock, small depth adjustments matter much less.
Limit order placement depth trades fill probability against adverse selection risk: resting at the touch fills faster but is more exposed to trading against better-informed order flow, while resting deeper waits longer for a fill but tends to capture a better price when it does fill.
Think of the queue at the touch as the front of a line that clears fast but where you're first to be handed whatever comes through the door — good or bad. Resting deeper is stepping back from the door: you see more before deciding whether the line is even worth joining.
Related concepts
Practice in interviews
Further reading
- Cartea, Jaimungal & Penalva, Algorithmic and High-Frequency Trading, ch. 6