You click buy on EURUSD at 1.0842. Half a second later a box appears: the price is now 1.0844, accept or reject. That is a requote, and the trade you thought you had never existed. On a quiet Tuesday afternoon it should be close to unheard of. In the first minute after a US payroll print it can happen on every click.
Requotes are not proof of dishonesty. They are what one specific order handling model does when it meets a fast market. The useful questions are how often they appear when the market is calm, whether they lean in one direction, and what the alternative model does with the same order.
What happens between the click and the fill
The order leaves the platform, crosses the internet to the broker's gateway, gets checked for margin, size and symbol permissions, and is then either filled from the broker's own book or routed out to a liquidity venue. Every hop costs time. A server sitting in the same London or New York data centre as its liquidity providers measures that round trip in single digit milliseconds. A trader on home broadband two continents away adds a couple of hundred milliseconds of pure network travel before any of the broker logic runs.
EURUSD can move a pip inside that window during a data release. So the quote your click was based on may already be history by the time the request lands. Every broker has to decide what to do about that gap, and the decision they made is visible in your trade history. If you want the full picture of who is on the other side of the fill, the difference between A book and B book routing matters more than the platform logo on the terminal.
Instant execution invites requotes, market execution invites slippage
Instant execution means the trader is asking to trade at a named price. If that price is still available, the fill matches the click exactly. If it is not, the server cannot fill at the requested level, so it comes back with a new price and waits for confirmation. That is the requote. It is a fair mechanism in theory, and a painful one in practice, because the confirmation window is often three seconds long and the market does not pause for it.
Market execution means the trader is asking to trade now, at whatever the book offers. There is no rejection and no confirmation dialogue. What you get instead is a fill that can be better or worse than the screen price, which traders call slippage. Most modern retail accounts run market execution, and stop losses behave the same way on both models: once triggered, a stop becomes a market order and takes the next available price.
A stop loss is a request, never a promise. On a gap open or a central bank surprise, the next available price can sit far below the level you set. This is the single most common reason a risk plan that looked tidy on paper fails on a Monday morning.
The numbers that describe execution quality
Execution quality is measurable, and the metrics are boring in the best way. Ask any broker or platform vendor for these five and see how quickly the answer arrives.
| Metric | What it measures | What a good answer looks like |
|---|---|---|
| Rejection rate | Share of orders returned unfilled, including requotes | Published per symbol group, and roughly flat outside news windows |
| Median execution time | Gateway receipt to fill confirmation | Quoted in milliseconds, with a percentile, not an average alone |
| Average slippage | Mean difference between requested and filled price | Small relative to the spread on that symbol |
| Slippage symmetry | How often slippage helps versus hurts the trader | Both directions occur, in similar proportions |
| Price improvement rate | Share of fills better than the requested price | A real, non-zero number that the broker can show you |
Symmetry is the tell
Slippage in a genuinely fast market is random with respect to your direction. Prices move both ways, so some fills land better than requested and some worse. When a trade history shows hundreds of negative slips and almost no positive ones, the randomness has been filtered somewhere. That filtering is what regulators mean when they write best execution rules and ask firms to prove that client orders were handled consistently.
The same logic applies to requotes. If the box only appears when the market has moved in your favour, the model has a thumb on the scale. If it appears equally when the move went against you, at least the mechanism is neutral, even if the experience is annoying.
Depth also matters. A one lot EURUSD order fills off the top of the book at almost any hour, because liquidity in EURUSD is deep enough to absorb it without moving the price. The same size in an exotic cross may have to walk down several price levels before it is complete, and the average fill will be worse than the top quote. That is not a broker failing. It is what thin depth costs, and no routing engine can invent liquidity that is not there.
Testing it yourself in two weeks
Open the account small and run a plain experiment. Place twenty market orders in calm hours, twenty in the London and New York overlap, and note the requested price against the fill price for each. Then repeat during a scheduled release you already follow from the news trading calendar. Log every rejection, every requote, and every difference between screen and fill.
Two weeks of that gives a real distribution instead of an anecdote. Traders who keep a trading journal with fill prices already have most of this data and never look at it. The pattern that matters is the shape: calm hours near perfect, news windows noisy in both directions. Calm hours already noisy is a different story, and worth acting on.
For firms building the other side of this, the fix is architectural rather than cosmetic. Put the matching layer close to the liquidity venue, keep the price feed and the order gateway on the same clock, and record every quote that was live when an order arrived so any dispute can be settled from the log instead of from memory. Platform latency is where most of the complaint volume in a support desk comes from, and it is cheaper to engineer away than to argue about.
"If a broker will not tell you their rejection rate, assume it is a number they are not proud of."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- A requote rejects the order and asks for a new confirmation; slippage fills it at a different price. Instant execution produces the first, market execution the second.
- Symmetry beats size: slippage that lands both for and against the trader is normal, slippage that only ever hurts is a routing question.
- Stop losses convert to market orders when triggered, so they carry no price guarantee through a gap or a data release.
- Twenty logged fills in calm hours and twenty in news hours tell you more about a broker than any review site.
Frequently Asked Questions
Does a requote always mean the broker is cheating?
No. A requote is what an instant execution model does when the price a trader clicked is no longer available. During a fast news print that is honest behaviour. The warning sign is requotes on quiet sessions, requotes that only appear when a position is in profit, or requotes that consistently move the price against the trader.
What is the difference between a requote and slippage?
A requote stops the order and asks the trader to accept a new price, so nothing is filled until a human clicks again. Slippage fills the order anyway, at a price different from the one requested. Instant execution accounts tend to produce requotes, market execution accounts tend to produce slippage.
Can limit orders avoid requotes?
A limit order cannot be filled worse than its price, so it removes the requote question for entries. The trade-off is that the order may never fill at all if price does not trade back to the level. Stop orders and stop losses stay exposed, because they convert into market orders once triggered.