Every VPS provider claims “ultra-low latency.” We decided to stop talking and start measuring. We ran controlled latency tests from a home PC on residential internet and from an FXVPS London (LD4) instance, connecting to the same broker servers, at the same time, using the same measurement tools. Here are the unedited results.
Test Methodology
We wanted numbers that any trader could reproduce, so we kept the setup simple and transparent.
Home PC setup:
- Location: Greater London area
- Connection 1: Virgin Media cable, 500 Mbps down / 50 Mbps up
- Connection 2: BT fiber (FTTP), 300 Mbps down / 50 Mbps up
- PC connected via Ethernet (no WiFi — we tested WiFi separately)
- Windows 11, MT4 build 1415
VPS setup:
- FXVPS London instance, located in Equinix LD4
- Core plan (1 dedicated vCPU, 2GB RAM)
- Windows Server 2022, MT4 build 1415
📊 Key Stat: Major forex brokers cluster in a handful of Equinix datacenters (LD4 London, NY4 New York, TY3 Tokyo). A VPS in the same facility achieves sub-millisecond latency because data travels meters, not continents.
What we measured:
- ICMP ping (100 packets, averaged)
- TCP connection establishment time (SYN to SYN-ACK)
- MT4 reported latency (bottom-right corner of the terminal)
- Traceroute hop count
- Packet loss over a 1-hour sustained test (1 packet/second)
All tests were conducted on a Tuesday between 14:00-16:00 UTC during active London/New York overlap — peak trading hours when networks are busiest.
The Results
| Metric | Home PC (Fiber) | Home PC (Cable) | FXVPS London (LD4) |
|---|---|---|---|
| Ping to IC Markets | 12-25ms | 18-45ms | 2.50ms |
| Ping to Pepperstone | 14-28ms | 20-50ms | 1.99ms |
| Ping to FTMO | 15-30ms | 22-48ms | 2.32ms |
| MT4 connection time | 200-500ms | 300-800ms | 10-30ms |
| Network hops | 12-18 | 15-22 | 1-3 |
| Packet loss (1hr test) | 0.1-0.5% | 0.2-1.0% | 0.00% |
The MT4 connection time is particularly telling. That is how long it takes your terminal to establish a fresh connection to the broker’s trade server. On a home connection, you are looking at a quarter to half a second. From LD4, it is nearly instant.
Why the Difference Is So Large
The raw numbers tell a story, but understanding why the gap exists matters more.
Home internet path: Your PC → home router → street cabinet → ISP local exchange → ISP backbone → internet exchange point (LINX) → broker’s datacenter network → broker’s trade server
That is 12-18 network hops. Every hop introduces latency (0.5-5ms each) and a chance for packet loss. Your ISP routes traffic for cost efficiency, not speed. The path your packets take to reach Equinix LD4 from your living room might go through Manchester before coming back to London.
📊 Key Stat: A home internet connection typically adds 50-200ms of latency to every trade. A colocated VPS in the same datacenter as your broker cuts that to under 5ms, directly improving fill prices on every order.
VPS in the same datacenter: VPS → top-of-rack switch → cross-connect → broker’s trade server
That is 1-3 hops. When your VPS sits in the same Equinix facility as your broker’s servers, traffic never touches the public internet. It travels over a direct fiber cross-connect within the building. The packets literally travel a few hundred meters of cable instead of dozens of kilometers.
Home internet is also shared bandwidth. Your ISP oversells capacity because not every household streams Netflix at the same time. But during peak hours, congestion spikes and your latency gets worse — exactly when markets are most active and execution speed matters most.
Datacenter connectivity is different. Dedicated fiber with guaranteed bandwidth, redundant paths, and no congestion. The connection quality does not degrade at 3pm when your neighbors get home from work. For practical advice on picking the datacenter closest to your broker, see tips for choosing the best VPS location for trading.
What This Means for Your Trading
Abstract millisecond numbers do not pay the bills. Let us translate this into money.
On a typical scalp, you enter and exit within minutes. The price moves fast. (For a deeper look at how latency affects high-frequency strategies, see scalping and high-frequency trading on a VPS.) Your order hits the broker’s server, gets routed to the liquidity provider, and gets filled at whatever price is available when it arrives. Every millisecond of delay means the price has moved slightly further from what you saw on your screen.
A 20ms latency disadvantage on a scalp translates to roughly 0.2-0.5 pips of additional slippage. That does not sound like much until you multiply it.
💡 Tip: The main “con” of a VPS — the monthly cost — is easily offset by the slippage savings, electricity savings, and protection against home internet outages. For active traders, a VPS is an investment, not an expense.
- 100 trades per month with 0.3 pip average extra slippage
- That is 30 pips lost to latency alone
- At $10 per pip on 1 standard lot: $300/month in hidden execution costs
- At $1 per pip on 0.1 lots: $30/month — still more than the VPS costs
This is not theoretical. If you track your actual fill prices against the price at the moment you clicked (or your EA triggered), you will see the slippage. Most traders blame the broker. Often it is their internet connection.
WiFi and Mobile Make It Worse
The numbers above used Ethernet connections. WiFi adds another layer of pain.
WiFi introduces 2-10ms of jitter on top of your base latency. Jitter means your ping is not a stable 15ms — it bounces between 12ms and 35ms unpredictably. One packet takes 12ms, the next takes 28ms, the next takes 15ms. Your EA cannot compensate for this randomness.
⚠️ Warning: High jitter (large gap between min and max latency) is as dangerous as high average latency. Inconsistent execution makes strategy performance unpredictable — look for both low average and low variance in your ping tests.
Mobile trading is worse still. 4G/5G connections typically show 50-200ms latency with significant jitter and periodic dropped connections during cell tower handoffs. Trading from your phone should be for monitoring and emergency closes, not active execution.
Traceroute: Seeing the Path Your Data Takes
We ran traceroutes from both locations to the same broker server. The home connection showed 16 hops:
1 192.168.1.1 (home router) 1ms
2 10.x.x.x (ISP local node) 4ms
3 core-router-lon (ISP backbone) 8ms
4 linx-peer (internet exchange) 10ms
...
16 broker-server (destination) 18msFrom the VPS:
1 10.x.x.x (datacenter switch) 0.1ms
2 cross-connect (direct fiber) 0.3ms
3 broker-server (destination) 1.9msFewer hops means fewer points of failure, lower latency, and zero exposure to public internet congestion.
The Verdict
A VPS is not just “faster internet.” It is a fundamentally different network topology. Your trading terminal sits inside the same facility as your broker’s servers, connected by direct fiber instead of routing through a dozen ISP nodes and public internet exchanges.
The difference between 2ms and 25ms might seem small in human terms. But your EA processes hundreds of ticks per second, and every order competes with other traders who might already be colocated. You are not just paying for speed — you are paying to remove an unnecessary disadvantage. For more on how latency improvements translate into concrete trading gains, read boosting trading speed with VPS.
Check Your Broker’s Latency
We publish real-time latency data for 195+ brokers from all FXVPS datacenter locations. Before you sign up, check what latency you would get to your specific broker from our broker latency page.
If you are losing money to slippage and cannot figure out why your fills are worse than expected, the answer might be sitting between your PC and your broker’s server. View our plans starting at $29/month — less than what most active traders lose to latency in a single week.