From the outside, placing a trade looks like a single event. You click buy, a position appears, and the platform reports a fill price. It feels atomic, the way sending a message feels atomic.
Behind that click is a chain of systems, relationships and decisions, and understanding roughly how it works explains most of the differences you notice between brokers.
(Disclosure: I am the founder and CEO of TabTrade. Some of what follows draws on how we have built our own infrastructure. I will keep this at the level of how the industry generally works, partly because the specifics are commercially sensitive at every broker, not just ours.)
Where the price comes from
The number on your screen did not originate with your broker.
Foreign exchange pricing starts with large institutional sources, principally banks operating globally, which quote continuously against one another. That institutional market is where price discovery genuinely happens.
Most brokers do not connect to those institutions individually. Instead they work with liquidity providers, which aggregate pricing from a range of institutional sources and pass a consolidated feed downstream. There are tiers of liquidity provider, just as there are tiers of regulator, and which tier a broker can access depends on its size, its capital, its relationships and its history.
Brokers then commonly aggregate again. Rather than relying on a single provider, a broker may take feeds from several and combine them, which improves resilience and typically tightens the resulting spread. That aggregated feed is what reaches your platform.
This is the first reason two brokers show different prices for the same instrument at the same moment. They are not looking at the same inputs.
The layer in the middle
Between the aggregated feed and your trading platform, most brokers run a risk management system. This is the part clients rarely hear about, and it does several jobs at once.
It monitors trading activity. Brokers have obligations to watch for market abuse and for clients attempting to exploit the system or the market, and that monitoring has to happen somewhere.
It manages pricing. A broker can apply markup or markdown to the feed it receives, which is how a standard account’s spread is constructed. Market conditions change constantly, and liquidity thins during volatility and during quieter sessions. A broker that wants to deliver a consistent client experience has to actively manage that feed rather than passing through whatever arrives.
It manages the broker’s own exposure. Not every trade is routed straight out. Brokers may warehouse some positions and hedge others externally, which is a normal part of running the business and is governed by internal risk policy.
The design of this layer is one of the biggest sources of variation between brokers, and one of the least visible.
Where your order goes
When you click buy, the order passes through that risk layer and, depending on the broker’s policy and the characteristics of the trade, is routed onward to a liquidity provider that takes it to market.
The platform you are looking at, whether that is MetaTrader 4 or 5, cTrader, DXtrade, Match-Trader or increasingly a charting front end like TradingView, is the interface. It is not the infrastructure. Most brokers license a platform rather than building one, because building genuine trading infrastructure from scratch is an enormous undertaking. What differs between brokers is everything sitting behind that shared front end.
Latency, and why location matters
Execution speed is partly a function of physical distance.
Serious brokers place their infrastructure in the major financial data centres where liquidity providers also sit, because proximity removes round-trip time that no amount of software optimisation can recover. Equinix NY4 in the New York area and LD5 in London are the two that come up most often in this industry.
This is also why algorithmic traders care so much about where their VPS sits. If your strategy opens and closes positions within seconds, the difference between being milliseconds from the broker’s servers and being on the other side of the world is not academic. It changes fills.
For a broker, fast execution is not only a service quality question. If your execution is slow, you become vulnerable to participants whose own pricing is faster than yours, because they know where the market is going before your system does. Speed is partly defensive.
Judge execution the way you should judge spreads
Here is where this connects to something I have written about before.
Brokers advertise execution speed the same way they advertise spreads, which is to say they quote the best number they have ever recorded. “Execution from 1 millisecond” may be perfectly true and tell you almost nothing, because it might describe a quiet moment on a liquid pair rather than the conditions under which you actually trade.
The meaningful figure is the average, measured over an extended period, across real market conditions. At TabTrade we track execution speed on that basis and have maintained an average under 30 milliseconds over a monthly window. I am quoting an average deliberately, for exactly the reason I set out in the piece on average versus minimum spreads: a best-case figure is a marketing artefact, an average is a description of what you will actually experience.
If a broker cannot tell you its averages, that is worth noticing.
Why this varies so much between brokers
Every component in this chain is a choice. Which liquidity providers, how many, how the feeds are aggregated, how the risk layer is configured, which data centres the infrastructure sits in, how the platform connects to everything else.
There is no standard configuration, which is why the same strategy can perform differently at two brokers advertising superficially identical conditions, and why a broker that suits one style of trading may be poorly matched to another.
It is also why testing beats reading. Open a demo account, put the platforms side by side, and watch the pricing and fills under the conditions you actually trade in. Fifteen minutes of that will tell you more than any comparison table.
Frequently Asked Questions
Does my broker take the other side of my trade?
It depends on the broker and the trade. Many brokers operate a hybrid model, routing some flow directly to liquidity providers and internalising or warehousing other positions according to their risk policy. This is normal industry practice and is generally described, at least in general terms, in the broker’s disclosure documents.
Why do two brokers show different prices at the same moment?
Because they are aggregating different liquidity sources and applying different pricing logic on top. Neither price is necessarily wrong. They are constructed from different inputs and managed by different systems.
Where should I put my VPS for algorithmic trading?
As close as possible to the data centre hosting your broker’s trading servers, which for most brokers means the New York or London facilities. Ask your broker directly where its servers are located rather than guessing, since brokers do migrate infrastructure.
Is faster execution always better for me?
Faster is generally better, but it is one factor among several. Consistent execution under volatile conditions matters more for most traders than a headline best-case speed, and execution quality also includes whether you receive price improvement when the market moves in your favour, not only how quickly you are filled.
