Latency vs jitter in trading, what they mean and how to measure both
The short answer
Latency is how long data takes to reach your broker and return. Jitter is how much that time changes. Measure both with 100 test pings, in a quiet hour and a busy one.
Latency is the delay you expect, jitter is the delay that surprises you
Latency is the usual time a piece of data takes to reach your broker and come back. Jitter is how much that time changes from one try to the next.
Why it matters. You can plan around a steady delay. A delay that jumps is the one that catches you out as you click.
Both are standard measures, defined by the internet standards body. The time there and back and how much it varies. So any test tool you use measures the same thing.
The table puts the two side by side, so you know which reading on a test shows which.
| Question | Latency | Jitter |
|---|---|---|
| What it describes | The usual time there and back | How much that time varies |
| How a test shows it | The average reading | The gap between the lowest and the highest |
| What sets it | Distance, and the path your data takes to your broker | Busy connections, and lost data that is sent again |
The standard notes that uneven delays make many real-time programs hard to run. Which costs you more depends on how you trade, so measure both.
Five steps give you your own numbers, quiet and busy
These steps show whether your delay comes from distance, your Wi-Fi or a busy path. Each one needs a different fix, so it pays to know which.
- 01
Get your broker’s server address
Your platform lists your broker’s trade server by name, and your orders go there. If that name does not answer, ask your broker’s support for one that does. The guide to measuring broker latency shows three ways to find it.
- 02
Send 100 test pings
A ping is a tiny check that times data going to your broker and back. Send 100 pings, because Windows sends 4 by default and can miss a rare slow reply. Write down the lowest, average and highest times, plus how many pings were lost.
- Windows
ping -n 100 broker-address- macOS and Linux
ping -c 100 broker-address
Linux also prints a deviation figure. The higher it is, the more the times vary (ping manual).
- 03
List the path your data takes
A traceroute lists each router between you and your broker, one per row. Routers may ignore these checks, so a gap in the middle proves no fault (mtr notes). Judge the path by the last row. That is the time to your broker.
- Windows
tracert broker-address- macOS and Linux
traceroute broker-address
- 04
Repeat in a busy session
Run the same commands when you normally trade hard, from the same computer and connection. A quiet hour tells you little about a busy one. The busy hour is when a delay costs you.
- 05
Change one thing at a time
To test your Wi-Fi, run the same test over a cable. If the gap between the lowest and highest time shrinks, the Wi-Fi was adding it. That is a fix in your own hands.
Your lowest time is distance, the rest is the path
The lowest time is your connection at its best, set by distance and link speed. Times above it point to busy links on the way to your broker (the standard). The gap between the highest and the lowest is a rough measure of jitter.
What it means for you. The lowest number is what distance costs you. The gap shows how much worse a bad moment can get.
Note. An example with made-up numbers
Say 100 test pings give a lowest of 20, an average of 24 and a highest of 140 milliseconds. The worst reply took seven times as long as the best. Repeat the test at another hour to see if the swings follow a busy time.
Find the row that matches your readings, so you know what is likely causing the delay and what to try next.
| You see | It suggests | Try next |
|---|---|---|
| Lowest, average and highest close together | A steady connection, with a delay that is mostly distance | Look at your broker’s execution mode and the spread |
| Average near the lowest, highest far above it | Occasional spikes, from a busy link or lost data | Note when the spikes happen, then repeat at another hour |
| Average well above the lowest, highest far above it | Busy links that come and go on the way to your broker | Repeat over a cable and at another hour |
| Steady when quiet, wide when busy | The delay rises with load, yours or your broker’s path | Repeat at the same busy hour on another day |
| Lost pings or timeouts | Data lost on the way, or a server that ignores test pings | Note when they happen, then repeat the test |
If the numbers stay steady and fills still surprise you, read the guide to requotes, slippage and disconnects. Testing a strategy instead? Read why backtests differ from live trading.
Distance sets a floor no one can remove, lost data causes the swings
This section explains where your delay comes from, so you know what anyone, us included, can change.
Light in a vacuum travels 299,792 km every second (NIST value). In fibre-optic cable it travels at about two-thirds of that (AFL Hyperscale).
Take a broker’s server 1,000 km away. There and back is 2,000 km. At 200 km every millisecond, that adds at least 10 milliseconds.
A real path is never shorter than a straight line, so this minimum only goes up. Latency is the minimum plus whatever the path adds.
What it means for you. No setting, proxy or faster internet plan can beat the minimum that distance sets.
Swings come from what happens on top of that minimum. Busy links make data wait in line. Lost data is sent again after a wait, and data must arrive in order (the standard), so everything behind it waits too.
Standard settings base that wait on the time there and back. It doubles each time the same piece of data is lost again (the standard). So the farther you are from your broker, the longer each wait.
Why it matters. This is the wait that delays price updates in a busy moment and holds up your next order.
You can fix some swings yourself, often for nothing
Start with the fixes in your own hands, because most cost little or nothing to try.
- Use a cable if the cable test was steadier than Wi-Fi. That removes the swings your Wi-Fi adds.
- Repeat the test at another hour, to see whether the swings follow a busy time.
- Send your internet provider the results, with their times, if the swings are between you and the internet. Dated readings give them something to act on.
- Ask your broker whether another trade server suits your location better. A nearer server lowers the floor itself.
Distance is different. Only a shorter path changes the minimum, and no setting on your computer does.
We recover lost data inside your broker’s city, not shorten the distance
A backtest runs on your own machine, so no internet link can lose a piece of data during a news release. In live trading, one lost piece holds up every price update and order behind it. Everything waits while your computer sends it again across the world.
With Traders Proxy, our server in your broker’s city opens its own, separate connection to your broker’s trade server.
If data drops on that local link, our server recovers it inside your broker’s city. It does not have to travel all the way back to you.
When data drops near your broker, your order waits on that short local link. Trading ports go first when a server is busy, and broker addresses stay warm in memory.
That helps only on the connection between our server and your broker. Data lost between you and our server is recovered between you and our server.
Our server keeps its waiting line for unsent data short on both connections. So new data doesn’t wait behind old data. That covers our server only, not your own computer or router.
You cannot paste four values into your platform while your chart is already frozen in a live trade. Set it up today while the market is calm. Then time it on your own screen during the next release.
We don’t promise lower ping. Distance sets the minimum. We make no claim about jitter. A test ping goes straight to the address and not through a proxy, so it cannot show what our server does.
So to judge us, time how long your platform takes to connect and show live prices, with and without us.
A VPS moves your whole platform to another computer, which can put it nearer your broker. We leave your platform where it is and change only where it connects. See Traders Proxy or a VPS for when each fits, and how Traders Proxy works for the full picture.
Check your broker’s terms first. If your broker prohibits proxies, don’t use one.
Common questions about latency and jitter
Frequently asked questions
What is a good latency for trading?
No single number is right, and no standard sets a cut-off. It depends on how far you are from your broker’s server and how you trade. Compare your quiet and busy readings with each other and with your own results.
Why does my speed test look fine when my platform feels slow?
A speed test measures how much data moves each second. Latency and jitter measure how long one small piece of data takes and how much that varies. A fast line can still have uneven delays.
Do test pings to my broker help?
Yes, over a long run. 100 pings show swings that 4 hide. They are still only a guide, because they are not your platform’s own traffic. A server that does not answer shows only timeouts.
Does Wi-Fi add jitter?
Test it. Run the same test over Wi-Fi and over a cable. If the gap between the lowest and highest time shrinks on the cable, your Wi-Fi was adding it.
Will Traders Proxy change my ping or jitter?
We don’t promise lower ping. Distance sets the minimum. We make no claim about jitter. Instead, our server opens a separate local connection in your broker’s city so lost data near your broker recovers inside that city.
What if I am far from my broker?
Distance sets a minimum that no network setting removes. A VPS nearer your broker can shorten the distance, and Traders Proxy does not. Our help covers only the connection between our server and your broker.
Field reports · Verified forum threads
Real forum reports, what our system does, and where it stops.
Every quote below links to the forum thread or vendor guide where it appeared. Underneath each quote, see which parts of our system work together on it and where our server stops.
cTrader slow or awkward for local-machine or tooling reasons, not network (UI stalls, Mac over remote desktop, Windows-only CLI, Open API apps)
“Unfortunately your description of your problem is not clear.”
Why a proxy does not change this. Traders Proxy doesn't change your computer's speed or the tools your platform offers.
Where it stops. Do not use these threads as proof for any site claim (two of the four cited links do not carry the quoted text).
The platform's proxy setting does not cover everything (WebRequest, MQL5.community, Market)
“Enable proxy server — allow the use of a proxy server when connecting to the trade server.”
Why a proxy does not change this. Only the trading app you set up uses Traders Proxy. Everything else stays on your normal internet.
- Only the app you set up is routedBoundary
Where it stops. Only the trading app you set up uses Traders Proxy. Expert Advisor web calls, custom DLLs and Market downloads may still use your normal internet.
Invalid account, Authorization failed and an empty broker server list
“I'm not able to search for other brokers in MT5 (Windows) Version: 5.00 build 4817 25 Jan 2025.”
Why a proxy does not change this. Traders Proxy forwards your platform’s traffic. It does not create, verify or fix your broker login.
Where it stops. If the Journal reports Invalid account or Authorization failed, your terminal reached the server and your login or server name needs checking.
Long-distance delay and undersea cable faults between regions
“Find a broker that uses cloudflare. I use VPN but from Europe. First traded with Generic Trader and ping was 200-300ms. then I changed to Ninjatrader and now my ping is 15-45ms depending on which serverlocation I use as VPN.”
Why a proxy does not change this. No claim. We don't shorten the distance between you and your broker's city, and we don't promise lower ping.
Where it stops. If your route to your broker’s city is long or damaged, our server cannot repair that cable fault and adds a step along the way.
Sources
- IETF standard 2681, a metric for delay there and backChecked
- IETF standard 3393, a metric for variation in delayChecked
- IETF standard 6298, computing the wait before lost data is sent againChecked
- IETF standard 9293, standard rules for ordered data deliveryChecked
- Microsoft, ping command referenceChecked
- Microsoft, tracert command referenceChecked
- Linux manual, ping(8), iputilsChecked
- Linux manual, traceroute(8)Checked
- mtr project README, probes and apparent lossChecked
- AFL Hyperscale, the speed of light in optical fibreChecked
- NIST, speed of light in vacuum (CODATA 2022)Checked