How Broker Choice Affects Trade Execution

When you press buy in MetaTrader, the order does not go “to the market” in any direct sense. It travels to your broker’s server, and what happens next depends entirely on how that broker is set up: the order may be filled from the broker’s own book, or routed onward to one or more liquidity providers— banks and non-bank market makers whose quotes make up the broker’s price feed.

Every stage of that path — the routing decision, the feed behind it, the physical distance between servers — is broker-specific. That is why broker choice is not a branding question but an execution question, and why it leaves measurable traces in your trade history.

Retail brokers are usually described with two broad labels. A dealing desk (market-making) broker can take the other side of client orders itself and manage the resulting exposure internally — hedging some of it with its own counterparties, netting the rest. An STP/ECN-style broker instead passes orders through to external liquidity providers or matches them in a pool, earning a commission or a small markup on the routed price. Many real brokers mix both models, routing some flow and internalizing the rest.

The clearest broker difference is how the transaction cost is packaged. The underlying interbank spread on a major pair is often a fraction of a pip; what you pay on top of it arrives in one of two shapes:

The only fair comparison is the all-in cost per trade, converted to pips or account currency. How spreads behave through the day — and why they widen around news — is covered in the spread guide; here the point is simply that the same cost can hide in two different places.

Between the moment your platform sends an order and the moment the broker fills it, price can move. Under market execution the order fills at whatever price is current when it arrives — the difference from the requested price is slippage, which can be negative or positive. Under instant execution the broker may instead reject the price and offer a requote, leaving you to accept the new price or cancel.

Fill policy differs too: whether stop orders trigger on bid or ask, how partial fills are handled, and how the server behaves in fast markets. Two brokers running the same nominal model can still fill the same stop order several tenths of a pip apart — small per trade, systematic across hundreds of trades.

A broker’s quotes are an aggregate of its particular liquidity providers. Different providers mean slightly different bids, asks and tick timing, which is why the “same” candle can print a different high or low at two brokers — and why a stop placed at an identical level can be touched at one broker and missed at the other.

Physical distance matters as well. If the broker’s trade server is in one city and your platform runs at home in another, each order crosses that distance before it can be filled, adding tens to hundreds of milliseconds in which price keeps moving. How that round-trip time translates into execution outcomes is covered in the latency guide.

Execution differences do not end at the fill. Swap rates— the overnight charge or credit on held positions — are set per broker per symbol, derived from interest-rate differentials plus each broker’s own adjustment. The same EUR/USD position held for ten nights can cost noticeably different amounts on two accounts, which matters a great deal to swing and carry-style strategies and not at all to day traders.

Each broker also publishes its own symbol specifications, and strategies inherit them silently:

• Contract size— “one lot” is not universal. A forex lot is typically 100,000 units, but an index or metal CFD can be 1 unit per lot at one broker and 10 at another, changing position-sizing math entirely.
• Stop level — the minimum distance, in points, at which stops and pending orders may be placed. A 20-point stop level can block orders that a tight-stop strategy places freely elsewhere.
• Minimum lot and lot step — coarser steps mean a fixed-percent risk model rounds to a different size, so even identical signals produce different exposure.

Put the pieces together and the divergence stops being mysterious. Consider one strategy run identically on two accounts for 200 one-lot EUR/USD trades — a scenario, not a claim about any real broker:

The 0.5-pip gap looks trivial, but across 200 trades at $10 per pip it is $1,000 of difference on identical signals. Sensitivity scales with trade size: a strategy averaging +3 pips gross per trade keeps 1.5 pips net on Account A and 2.0 pips on Account B — a third more — while a strategy averaging +40 pips barely notices either cost base. The direction could just as easily be reversed; the point is the size of the effect, not which model wins. You can test how a half-pip of cost moves your own numbers in the Trading Expectancy Calculator.

Marketing pages cannot settle any of this, but your executed trades can. Every fill in a MetaTrader history carries evidence of execution quality, and reviewing it per symbol, per session and per order type turns “my broker feels slow” into numbers:

• Average spread paid— the spread recorded at your entries, broken down by symbol and hour of day, rather than the headline “from 0.0 pips”.
• Slippage — requested versus filled price on market and stop orders, separated into quiet hours and news windows, entries and exits.
• Requote and rejection frequency — how often orders failed to fill at the requested price on instant-execution accounts.
• Commission and swap per lot — totals from the history, converted to a pips-per-trade figure you can set against your average win.

Tracked over a few hundred trades on your own synced accounts, these figures describe the execution you actually receive — the only version of the question that matters for your results.