Home › Learn › Guides

Sub-Millisecond Forex Execution: What's Realistic for Retail?

Updated 14 July 2026 · 9 min read · PipTax education

Sub-millisecond forex execution is a phrase you'll see thrown around by VPS providers and "low latency" trading tool sellers, promising institutional-grade speed to ordinary retail accounts. The honest answer is: sub-millisecond latency to a matching engine is achievable in specific setups, but it rarely changes your trading outcomes the way the marketing implies — because spread, bridge processing, and order handling dominate what actually happens to your fills.

What Sub-Millisecond Actually Means

In institutional trading, sub-millisecond usually refers to the round-trip network time between a server and an exchange's or liquidity venue's matching engine — measured in microseconds by firms co-located in the same data centre as that engine. This is real, but it's a narrow slice of the whole execution chain.

For retail forex, here's what's actually happening between your click and your fill:

• Your terminal to your VPS (if you use one) — typically 1-20ms depending on location
• VPS to broker's trade server — this is the part that can genuinely be sub-millisecond if co-located
• Broker's bridge/aggregation layer — often adds several milliseconds as it checks liquidity, pricing, and risk
• Broker to liquidity provider(s) — another hop, more milliseconds
• Price/slippage checks and order confirmation back to you

So even if your VPS-to-broker leg is genuinely sub-millisecond, the total round trip a retail order experiences is almost always several milliseconds to tens of milliseconds — and that's normal, not a failure. Anyone claiming your entire retail trade lifecycle is sub-millisecond is being imprecise at best.

Why Proximity to the Server Still Matters

Physical distance to the server your broker's trade server sits in is the single biggest lever retail traders can actually pull. Many FX venues, prime brokers, and liquidity providers co-locate in a handful of major data centres — Equinix LD4 in Slough is one of the most important in Europe, alongside NY4 in New Jersey and TY3 in Tokyo.

Practical takeaways:

• If your broker's MT4/MT5 servers are hosted near LD4, a VPS in the same facility or metro area (London) will shave meaningful, measurable milliseconds off your VPS-to-broker leg compared with a VPS in another country.
• Pepperstone and IG both publish server location details in their platform/server lists — check which data centre your specific account's trade server actually sits in before choosing a VPS location.
• Don't assume "London VPS" automatically means "same building as your broker's server." Confirm with a latency test (below), not guesswork.

How to Measure Your Real VPS-to-Broker Latency

Don't take anyone's word for your latency — measure it yourself. This takes ten minutes and no special tools.

Using MT4/MT5: 1. Open the Journal and Experts tabs after connecting. 2. Right-click a chart → Trading → or check Tools > Options > Server ping shown in the bottom-right status bar of the platform — MT4/MT5 display a live ping in milliseconds to your trade server. 3. Watch it over a few hours, not just once — spikes during news events matter more than the average.

Using Windows tools on your VPS: - ping [broker-server-address] — gives round-trip time (RTT) in milliseconds; run 50-100 pings and look at the average and worst-case, not just the best. - tracert [broker-server-address] — shows every network hop your data takes, useful for spotting an inefficient route (e.g. traffic detouring through an unexpected country).

A realistic good result for a well-placed VPS is single-digit to low double-digit milliseconds. If you're seeing 100ms+, your VPS location or route is wrong for that broker.

Setting Up a Reliable 24/7 VPS for MetaTrader

A forex VPS setup only helps if it's actually running your terminal reliably around the clock. Key steps:

• Choose a location near your broker's server, confirmed by the ping test above, not just "closest to me."
• Install MT4/MT5 directly on the VPS (not just remote-controlling your home PC) so it keeps running when your laptop is off.
• Enable auto-start: set MetaTrader to launch on Windows startup and auto-login to your account, so a VPS reboot doesn't leave you flat.
• Configure Expert Advisor auto-start settings if you run automated strategies — enable "Allow live trading" and auto-attach EAs to charts on launch.
• Set the VPS to never sleep — disable Windows sleep/hibernate settings entirely, since a "sleeping" VPS is a VPS that isn't trading.

RDP and Firewall Hardening You Shouldn't Skip

A VPS running your trading account is a target — treat it like one.

• Change the default RDP port from 3389 to reduce automated scanning attempts.
• Use a strong, unique password and enable two-factor authentication on the VPS provider account itself, not just Windows login.
• Restrict RDP access by IP where your provider allows it, so only your own connection can reach the remote desktop.
• Keep Windows and MetaTrader updated — unpatched VPS instances are a common attack vector.
• Enable the Windows Firewall and only open the ports MetaTrader and your VPS provider's tools actually need.
• Set up email/SMS alerts for failed login attempts if your provider supports it.

None of this is optional if you're leaving real capital and live trades connected 24/7.

Latency vs Spread: Where Your Money Actually Goes

This is the part low-latency marketing conveniently skips: for the vast majority of retail strategies, spread, commission, and slippage cost you vastly more over time than a few milliseconds of latency ever will. A trader obsessing over shaving 5ms off VPS ping while trading on a wide, uncompetitive spread has the priorities backwards.

| Factor | Typical impact on retail outcomes | |---|---| | Spread/commission | Direct cost every single trade, compounds over volume | | Slippage on entries/exits | Variable, worse in fast markets, moderate impact | | VPS-to-broker latency | Small, mostly matters for scalping/high-frequency styles | | Broker bridge/execution quality | Often bigger factor than raw ping |

Where latency genuinely matters is scalping and some automated strategies reacting to fast price moves — there, a few milliseconds can affect fill quality at the margin. For swing and position trading, it's close to irrelevant.

Getting Sub-Millisecond Execution in Perspective

Sub-millisecond forex execution is a real, measurable thing at the network layer — but for retail traders it's one small piece of a longer chain that includes your broker's bridge, liquidity sourcing, and your own spread and commission costs. Spend your effort where it compounds: pick a VPS location backed by an actual ping test against your broker's server, harden it properly, and then turn your attention to the costs that hit every single trade. Run your own numbers with PipTax's [cost audit tool](/audit.html), compare execution details on the [brokers page](/brokers/index.html), and see how spread and commission compound over time with the [cost-impact calculator](/cost-impact.html).

Keep going: Audit Index Cost Impact Methodology