Speed and Performance: Copy Trading Measured in Milliseconds

Speed is one of the first things people compare when choosing multi-account or copy-trading software. Most of what you can read about it is a claim with nothing behind it.

This page takes the opposite approach. Every number below is measured from live production traffic, and every one of them is something you can check yourself from your own screen, down to the millisecond.

What we measured, and what it means

What	Measured result	What it covers
Time our platform adds to a copied order	Under 25 milliseconds (typical)	From the moment we decide to copy, to the moment the child order leaves for the broker
Our share of a single order placement	About 20 milliseconds (average)	Our own processing, excluding your broker
Full round trip to the broker, fastest network	9 out of 10 orders under 100 milliseconds	Order sent to the broker, and the broker’s confirmation received
A complete order placement, end to end	100 to 200 milliseconds (typical)	Everything: our processing plus your broker’s response
Turning one master order into all its child orders	About 22 milliseconds for 20 to 49 accounts	Our own work: first child order created and sent on, to last
Getting every one of those orders to the brokers	About 170 milliseconds for 20 to 49 accounts	First order to reach a broker, to last
PMS copy (you place the order here)	Almost no lag	No waiting for a broker to tell us about a new order
Master-child copy start	Under 0.2 seconds on high-performance brokers	How long your broker takes to notify us of the new order

Two things are worth reading twice.

Our own share is small, and it is the only part we control. In a 150-millisecond order, roughly 20 milliseconds is us and the rest is your broker and the network in between. That is why this page separates the two everywhere, and why the app does too.

Adding accounts does not multiply the time. That is the direct result of how the platform is built, and it is the subject of the next section.

Every account is processed in parallel, never one after another

This is the single most important thing to understand about how the platform performs.

When an action affects many accounts, those accounts are worked on together, not queued up behind one another. We do not send an order to the first account, wait for that broker to answer, and only then move to the second. Each account’s order is prepared and sent independently, so a slow broker on one account does not hold up the rest.

That applies across the platform:

• Master-child copy trading — when your master order is copied, every child account’s order goes out together. A master with 40 children is not 40 times slower than a master with one.
• PMS copy (placing one order into many accounts) — you tick the accounts or groups you want and click once. Every selected account is processed together; we never wait for one broker before starting the next.
• Bulk modify, cancel and square-off — the same. Selecting 30 orders to cancel does not mean 30 waits in a queue.
• Square-off portfolio — every open position is closed at the same time, not drained one position after another.
• Monitoring many accounts — your positions, holdings, margins and orders are fetched from all of your accounts at the same time, which is why a large book still loads as one screen instead of account by account.
• Multiple master orders at once — if your strategy fires several orders together, they are copied in parallel too. You can switch a master account to strict one-at-a-time copying if your strategy depends on a guaranteed order sequence; parallel is the default because it is faster.

Why it matters: a broker call takes roughly a tenth of a second. Handled one after another, 40 accounts would mean several seconds between your first account and your last — a real difference in a fast market. Handled together, the gap is measured in milliseconds.

Copy trading speed, measured across thousands of orders

Here is the direct evidence, taken from live production over the last 30 days, across more than 27,000 orders that went into 5 or more accounts.

Sending one order into many accounts happens in two steps, and we publish both — because only the first one is fully ours:

• Our own work — one order becomes many, and each one is sent onward. This part is entirely ours.
• Reaching the brokers — each of those orders is then a separate request to a separate broker. This step carries your brokers’ own speed with it.

Both numbers are the gap between the first account and the last, so they show how tightly your accounts are kept together.

Accounts in one order	Our own work, all done within	All orders reach the brokers within
5 to 9	2 milliseconds	9 milliseconds
10 to 19	10 milliseconds	40 milliseconds
20 to 49	22 milliseconds	170 milliseconds

Typical (median) values. Nine out of ten finish within roughly 10, 27 and 52 milliseconds for our own work, and 190 milliseconds, 0.5 seconds and 1.9 seconds for reaching the brokers.

So a trader with 27 accounts has all 27 orders created inside about 22 milliseconds, and all 27 sitting at their brokers inside about a fifth of a second. Sending them one at a time would take around four seconds.

The two copy methods have different starting points

Both methods copy in parallel once copying starts. What differs is when copying can start:

• PMS copy — you place the order in our terminal, so copying begins immediately. There is nothing to wait for. This is the faster of the two.
• Master-child copy — your order is placed at your broker (or by your algo), and your broker must tell us it exists before we can copy it. On high-performance brokers that notification arrives in under 0.2 seconds. On slower ones it takes 0.5 to 4 seconds. That wait is at your broker, not with us — which is why our documentation publishes which brokers are fast and which are not, instead of hiding it.

If speed is your first priority and you place orders by hand, use PMS copy. If you must copy from a broker terminal or an algo, choose your master account’s broker from the high-performance list.

Millisecond-level timing you can check yourself

Speed claims are easy to make. We would rather show you the measurement.

Every order on the platform carries its own timing record, accurate to the millisecond, and you can open it whenever you want:

1. The Copy Log’s Timing Statistics — for copied orders. Open a master order’s Copy Log (menu: Trading → Portfolio → Orders tab → Copy Log). The Timing Statistics section lays out the whole copy as a timeline: when we received the master order, when each child order was sent, and when each broker confirmed it. Every child account gets its own row, with separate Our system and Broker system columns and a total. Click the help icon in any cell for the stage-by-stage breakdown behind that number.

This is the part that matters: it tells you where the time went. If the Broker system column is large, the delay was at your broker. If our column is large, it is ours to answer for.

2. The activity log — for every order, including API orders. Open Tools → Activity. Every entry is timestamped to the millisecond. An order produces two entries — the moment we received it, and the moment your broker replied with an order id. Subtract one from the other and you have the exact time that order took. For example, 10:15:32.118 and 10:15:32.267 means the order took 149 milliseconds. This works for orders placed by hand, by your alerts, and by your own program through our API.

3. The result window’s Timing panel — for bulk actions. When you place, modify, cancel or square off across many accounts, the result window shows the total time the whole action took, alongside the per-order broker times.

Nothing here is a summary we prepared for marketing. It is the same measurement data we use ourselves when a user reports a slow order.

Where the rest of the time goes: your broker and your network

Once our share is 20 to 25 milliseconds, the honest conversation is about the other 80 to 90 per cent. Two things decide it.

Your broker. Broker APIs differ enormously — in how fast they accept an order, and in how quickly they announce a fill. Some brokers push an update the instant something happens; others must be asked repeatedly. We publish a full broker API comparison covering exactly this, including how each broker sends order updates, so you can choose with the facts in front of you.

Your network path. Some brokers only allow API access from a fixed IP address. On the older IPv4 network, providing a fixed address needs extra routing, and that routing adds delay to every order. On the newer IPv6 network it does not. Measured in production, this is the single biggest difference between two otherwise identical accounts:

Static IP option	Measured order round trip to the broker
IPv6 (fastest, ₹100/month)	9 out of 10 orders under 100 milliseconds
Bring your own IPv4	Close behind
Premium managed IPv4 (₹299/month)	Close behind
Standard IPv4 from our pool (₹200/month)	The budget choice — its shared routing adds noticeable delay to each order

If your broker supports IPv6, take it. If your broker is IPv4-only and speed matters to you, the premium option is the one to pick. See pricing for the full comparison.

What makes the platform fast

• Work is done in parallel, everywhere it can be. Covered above — it is the design decision that matters most.
• It runs in the cloud, close to the market. The platform runs on Google Cloud in its Mumbai, India region, so the distance between us and your broker is short. Nothing runs on your computer, so nothing slows down when your machine does, and there is no PC or rented server of yours in the path.
• Nothing to install, nothing to fall behind. Improvements ship almost daily and are installed for everyone at once, so every user is always on the newest and fastest version.
• You can choose your network path. The static IP options above let you buy speed where your broker forces a fixed address.
• We measure it continuously. Per-order timing is recorded for every copied order — which is how the numbers on this page exist at all, and how a slow order can be investigated instead of argued about.

How to run a fair speed test

If you want to test this yourself, a few things make the difference between a real measurement and a misleading one.

1. Test during live market hours. Outside market hours the system deliberately works at a reduced pace, and brokers behave differently. Results from a closed market mean nothing.
2. Skip the first order of the day. The first order for an account includes a one-time session setup with the broker. Measure from the second order onwards.
3. Use the timing report, not a stopwatch. Your own network, your browser and your screen refresh are all in the path of what you can see. The timing report measures the order itself.
4. Average over at least 100 orders. A single order tells you very little; brokers vary minute to minute, and market open is the busiest time of the day for all of them.
5. Compare the right timestamps. On a broker terminal, the time shown against an order is usually the last modified time, not the time it was created. Comparing that with a copy time will always look wrong. Use the created time, or use our activity log.

More detail on all of this is in Slow performance and in the copy-trading documentation.