Direct answer
Last Look matters in forex because it can change the outcome of an order after it is submitted. Instead of every request being immediately and irrevocably accepted, a venue may perform a brief “review” and then either confirm, amend, or reject the execution based on conditions at the moment of confirmation. This can affect practical decisions such as how you interpret fill rates, effective pricing versus quoted pricing, and how you evaluate past performance.
Mechanism and definition
Last Look is a trading/quote handling process where an incoming execution request is not guaranteed to be filled exactly as first seen. The venue may apply timing and market checks—often related to whether the quoted price or executable conditions are still acceptable during a small window.
A simple way to separate concepts is:
- Stable mechanics: the presence of a review step that can lead to acceptance or rejection.
- Variable conditions: the market state (price movement and liquidity), and the venue’s internal policy (how it decides whether an execution remains acceptable).
This review step matters because it changes what “execution” means. Two orders that look similar from the client side can produce different outcomes if the acceptance decision differs during that review window.
Scenario: why outcomes can differ
Assume you place a limit order with the expectation that it will fill if the market reaches your price. If a venue uses Last Look, then even if the market reaches or crosses your level, the venue may still reject or delay confirmation when conditions change within the review window. The practical effect is that you may observe:
- fewer fills than a no-review model would predict,
- different effective prices than the ones you assumed when placing the order,
- a mismatch between backtested assumptions and live results.
Evidence or example (with explicit assumptions)
Consider a hypothetical comparison:
- Assumption A: You measure quotes and decide the order should fill at a quoted level.
- Assumption B: In reality, the venue performs a brief conditional review.
- Assumption C: During that window, the market moves slightly or liquidity thins.
Under these assumptions, acceptance becomes conditional, so your observed fill probability is lower and your realized execution terms can differ. Importantly, this does not require any promised “better” or “worse” outcome in general; it only means the execution path can diverge from what a model assuming immediate acceptance would expect.
A useful verification angle is to compare expected fill conditions (what your system predicts under immediate acceptance) with observed outcomes (what actually happened). If the gap persists across changing conditions, that gap is often a sign that execution control steps—such as Last Look—are influencing acceptance decisions.
Limitations and risks, plus what you can independently verify
The main limitation is that Last Look implementations and details are not universal. Without the specific execution policy and how acceptance decisions are determined, you cannot reliably infer the exact mechanism from results alone.
Common failure modes in interpretation include:
- Treating realized fills as deterministic: you may mistakenly assume that reaching a price level guarantees acceptance.
- Ignoring timing and costs: latency, spreads, and transaction costs can change effective outcomes and can be mistaken for acceptance effects.
- Overgeneralizing from historical relationships: past fill behavior does not ensure future behavior under different volatility or liquidity.
To independently verify what matters for your use case, focus on concrete, non-promotional evidence: collect execution records, compute fill rates by conditions (for example, different volatility regimes), and document the assumptions in your evaluation. Because results vary with market conditions and jurisdiction, keep your conclusions conditional rather than predictive.
If you want a deeper next step, you can examine how worked execution examples translate request timing into realized outcomes, and how common interpretation mistakes distort the comparison between quoted and executed results.