[Development Log] How the choice of broker affects EA performance? Explaining differences in execution, swap, and rollover specifications

When operating FX automated trading (EA), many people tend to focus on the inherent advantage of the logic itself. However, even with the same logic, performance can change depending on which broker you run it with. This time, we’ll organize the mechanisms and key points to check on how “broker selection” affects EA performance.

■ Differences in brokers are differences in the foundation of operation

Even with the same EA, the same currency pair, and the same settings, results can vary if the broker is different. The reason is that each broker has differences in the “foundation” aspects such as execution speed, spreads, swap, and rollover handling. Even if the logic—the blueprint—is the same, the foundation on which it is executed can cause the final results to differ slightly. Underestimating this foundation can lead to a situation where “the logic should be superior, but for some reason the performance doesn’t improve.”

In particular, people who use multiple accounts or brokers tend to overlook these differences. It is not uncommon to think you are running the same EA on different accounts, but in reality each is operating under different conditions.

■ The impact of variation in order execution speed

First and foremost is the “execution speed.” Even with the same order, execution speed and the amount of price slippage (deviation from the expected price) differ by broker. For logic that enters or exits at times of rapid market changes, this difference in execution speed tends to translate directly into performance differences.

In particular, scalping-type or news-event-driven logics can be strongly affected by execution speed differences. Conversely, long-term swing-type logic is less affected by small differences in execution speed. Depending on the type of logic, the degree to which execution speed matters will vary. Additionally, the frequency of order rejection (re-quotes) and how wide spreads widen at economic data releases are points that can easily differ by broker. Since these aspects are hard to see on a demo account, verifying them with a small live account is a reliable method.

■ Differences in swap and rollover specifications

Next, the swap and rollover specifications are often overlooked. Swap points incurred when holding a position overnight vary in level by broker, and the timing and rules for accrual are also subtly different.

For swing- or trend-following logics that hold positions for long periods, these swap differences can have a non-negligible impact on final profit and loss. Even with the same logic run over the same period, the swap levels alone can create differences in total performance. Additionally, how weekend rollover is handled (how Saturday-Sunday swaps are calculated) varies by broker, so it’s worth confirming in advance when operating logics that assume long-term holding. Negative swap accumulating larger than expected can be especially painful for long-term operators.

■ Broker-specific factors that are hard to see in the validation stage

The tricky part is that these broker factors are not easy to see during backtesting. Backtests are generally based on “ideal execution,” and do not reflect the quirks and constraints of actual brokers.

Therefore, good validation results do not guarantee the same performance in real trading. To properly evaluate the advantage of the logic, it is necessary to consider the premise of “which broker and under what execution environment will it run.” Like spreads, slippage, and latency, environmental factors outside the logic are often treated as “errors” in the validation stage. However, once real trading begins, the accumulation of these errors can manifest as performance differences that cannot be explained by the logic alone.

■ Points to consider when choosing a broker

With what has been described so far, the key points to consider when selecting a broker can be summarized as follows.

・Assess whether you should prioritize execution speed or swap values based on the type of logic you use (scalping vs. swing)
・If possible, run the same logic in demo environments of multiple brokers in parallel and compare actual performance differences
・Check swap and rollover specifications not only from official information but also from actual accrual history to ensure there are no discrepancies with your expectations
・When switching brokers, don’t transfer all funds at once. Start with a small lot in parallel, confirm the performance difference, and then gradually shift the weight

“If the logic is excellent, it should produce the same results no matter where it runs” is not that simple. The differences in the broker foundation gradually affect performance, so be mindful. Especially if you are about to start full-scale EA operation, it is worth dedicating as much time to broker selection as to logic selection.

■ Summary: Test the logic and the broker together

We explained how broker selection influences the performance of the logic. Differences in execution speed, swap, and rollover may seem subtle, but they reliably affect long-term results.

Test the broker environment with the same care as testing the logic. By adopting such a perspective, you can reduce the gaps you may experience in actual operation.

? If you have questions about markets or logic, Semura_Lab Yorozu Consultations also accepts inquiries.
https://www.gogojungle.co.jp/mypage/community/485