Price-Time Priority and Order Matching
When several orders sit at the same price, the exchange needs a rule for who trades first. The dominant rule is price-time priority:
- Better price wins. Higher bids and lower offers execute first.
- At the same price, earlier wins. Orders are a FIFO queue per price level.
Both halves are incentives. The first rewards competing on price, which tightens spreads. The second rewards posting early, which encourages resting liquidity rather than everyone waiting.
Queue position is real value
Consider a bid at 99.98 with 500 shares ahead of you and 200 behind. A seller arrives with 400 shares: the 500 ahead absorb it entirely and you get nothing. Only after that entire queue clears do you trade.
- 1500 shares posted at 09:31:04, ahead of you in the queuefills first
- 2Your 300 shares posted at 09:31:22, and untouched by a 400-share seller
- 3200 shares posted at 09:33:10, behind you
A 400-share seller clears the top rung and stops. Everything below it waits, whatever the price on the screen says.
This has direct consequences for how market makers behave.
Getting in early matters. Being first in a queue at a price you like is worth a great deal, because you fill before anyone else at that level.
Cancelling is expensive. Modify an order and you lose your place, going to the back. So a market maker who wants to adjust faces a real cost, and this is one reason quotes are stickier than a naive model would predict.
Sometimes the better price is worth it. Jumping ahead by a tick sacrifices a penny per share but gains priority over the whole queue. Whether that trade is worth it depends on how likely you are to fill from your current position, which is a genuine optimisation traders make constantly.
Queue position is an asset. It is earned by posting early and destroyed by cancelling, which is why market makers think carefully before adjusting a quote they are happy with.
The matching engine
When a buy and sell price overlap, the engine matches them automatically, executing at the price of the order that was resting rather than the incoming one. The resting order set the terms; the arriving order accepted them.
Modern engines process millions of messages per second with deterministic ordering, which matters for fairness: given the same message sequence, the outcome must be identical every time.
Other priority models
Price-time is dominant but not universal.
Pro-rata, common in some futures markets, allocates fills proportionally to order size at a price. This rewards posting size rather than posting early, and it produces very different behaviour: participants quote much larger sizes to secure a bigger share of each fill, sometimes far beyond what they actually want to trade.
Size-time hybrids blend the two, giving some allocation to the top of the queue and distributing the rest pro-rata.
Knowing which model an exchange uses is not trivia. It determines the optimal quoting strategy entirely, and a strategy tuned for price-time will behave badly on a pro-rata venue.
Before trading a new venue, check its priority rules, tick size and fee structure. These three determine the mechanics of a quoting strategy more than the instrument itself does.
The interview angle
A common question is whether to join the queue at the current best bid or improve by a tick. The answer requires weighing the probability of filling from your queue position against the cost of the extra tick, and the expected adverse selection either way. There is no single right answer, and interviewers are listening for whether you identify those three factors at all.
The third factor is the one candidates drop, and it is the one that makes this a strategic problem rather than an arithmetic one. Improving by a tick raises your fill probability, but the fills you gain are disproportionately the ones you least wanted, so the value has to be computed conditional on trading rather than unconditionally. Expected value against an optimiser is that correction in isolation.