How to Benchmark FX Execution Quality: Slippage, Rejects, Requotes

Best execution is a habit, not a one-trade verdict

When your FX fill lands a few pips away from the price on screen, it is easy to feel either lucky or cheated. But a single trade rarely proves much. Execution quality is a pattern. It shows up when you compare dozens of fills against a fair reference price and notice how often you are filled, re-quoted, rejected or quietly improved.

Regulators treat this seriously. The UK Financial Conduct Authority has found firms failing to understand and monitor best execution properly, putting clients’ interests at risk. That does not mean every disappointing fill is a scandal. It means execution should not be an opaque black box. You can build a simple measurement habit using the same broad logic institutions apply, but on a scale that fits a normal trading account.

Start with a fair reference price

To judge any FX execution, you need a neutral yardstick. One useful reference is the fair market mid at the time you submitted the order: the midpoint of the best available bid and ask. If you compare your achieved price with that mid, you can see whether your fill was skewed in one direction. Some transaction-cost analysis treats the arrival price as the median one-second mid-quote at the moment the order starts, because a single flickering tick can be misleading.

For orders that take minutes or hours to complete, a second useful benchmark is the volume-weighted average price, or VWAP. VWAP tells you the average market price adjusted for traded volume over that period. If your overall order beats the fair mid or VWAP after fees, execution may be working well even if one individual fill looked ugly.

Slippage has a direction and a story

Slippage happens when a market order fills automatically at the next available price, without asking for confirmation. Negative slippage means the fill is worse than the price you saw. Positive slippage means the fill is better. Many traders only notice negative slippage, but a fair execution record tracks both.

One useful health check is the positive slippage ratio: the number of positive slippage events divided by the total number of slippage events. It will not be exactly balanced, especially in fast markets or with larger orders. But if a long sample shows almost no positive slippage while negative slippage is common, that is worth investigating. Execution quality is about distribution over time, not about avoiding every small drift.

Requotes are not the same as a bad fill

A requote is a distinct execution event. In an instant execution model, your order is tied to a requested price. If the server price has moved beyond your permitted deviation, the broker declines that price and offers you a new one. Only then do you choose whether to accept or cancel. The requested price, server price and allowed deviation can be expressed as a simple gap: if the absolute difference between the server price and the requested price is greater than the allowed deviation, the platform returns a requote.

For example, imagine you send a buy order at 1.25000 with a permitted deviation of 0.00005. By the time the order reaches the broker server, the executable ask is 1.25009. The difference is 0.00009, which is larger than your allowed 0.00005. The order fails the price check and a requote appears. That is not automatically misconduct; it is how fixed-price order models protect against stale prices.

The practical dial is your maximum deviation or slippage tolerance. A narrow setting protects price but increases requotes. A wider setting reduces requotes but allows more drift. The more revealing issue is asymmetry. If requotes cluster mainly when the market moves in your favour, while orders move freely when the market moves against you, that is a much sharper warning sign than the requote itself.

Last look is the market’s brief “let me check” pause

Last look is a short validation window used by some liquidity providers in over-the-counter FX. After your order arrives, the provider has a few milliseconds to accept the trade at the quoted price, reject it, or offer a requote if the market has moved. It exists largely to protect liquidity providers from latency, stale prices and sudden volatility. In calmer conditions it may be barely noticeable. Around major economic news, it becomes far more visible.

The length of that hold time matters. A shorter last look window tends to produce fewer rejections, but it leaves the liquidity provider carrying more price risk. A longer window reduces provider risk but can increase rejections and hold times for the person waiting on the other side. That is why fill timestamps and rejection rates are worth recording, not just the headline spread.

Last look is not inherently good or bad. It becomes a problem when it is hidden, unevenly applied, or used to act on information from your trade request. The Global FX Code’s working group has warned against practices such as pre-hedging during the window, unfair asymmetry, or artificially stretching the hold period. No-last-look execution is possible, but it often comes with wider spreads or stricter order limits because providers must commit before seeing how the market moves. The choice is usually about where the risk sits, not about a magical guarantee of better prices.

How to build a simple FX execution kit

You do not need an institutional platform to track the essentials. Keep a log that includes the time you submitted each order, the reference mid or arrival price, the final fill price, fees, whether the order slipped, was requoted, rejected or accepted at the quoted price, and how long any hold time lasted.

From that log, calculate a handful of simple measures: slippage from the fair mid in percentage or pips, rejection rate, requote rate, average last look hold time where visible, positive slippage ratio, and a rough requote bias between client-favourable and client-adverse events. Then compare the same numbers in calm London or New York hours versus around major economic releases. FX execution is heavily shaped by conditions, so a single snapshot can mislead.

Match the execution model to your strategy

Execution type is not a one-size-fits-all choice. A slower swing trader may prefer a price-committed instant execution model, because a requote gives them a chance to decline a stale price. A news scalper or algorithmic system may struggle with repeated manual confirmation; the same price movement often appears as automatic slippage in a market execution model instead of a requote. None of this is a promise of perfect fills, and brokers differ by region, asset class and liquidity source.

The goal is not to banish slippage, rejects or requotes entirely. The goal is to know what you are actually experiencing, compare it against a fair reference, and ask better questions about whether the execution pattern is consistent, disclosed and reasonably balanced.