Skip to content

14-day trial · Founding Rate live.

How Copilink measures trade copier latency

Real-world benchmarks on standard hardware

TRANSPARENCY

Overview

We publish product-specific internal processing benchmarks, measured in real Copilink workflows on desktop hardware. Copilink Standalone and Copilink for NinjaTrader 8 are separate products with different integrations, so their results are measured and reported separately.

These benchmarks describe the time Copilink spends processing a defined event inside the application. They do not measure broker, network, exchange, or end-to-end fill latency, and they do not guarantee matching fills or P&L across accounts.

Copilink Standalone benchmark

Copilink Standalone averaged ~5 ms of internal follower-order event processing. The reported average is 4.91 ms, rounded for readability.

  • Measurement window: Five Copilink sessions recorded September 9-11, 2026, representing about 5 hours and 14 minutes of runtime.
  • Observed activity: 2,171 follower order events and 585 leader order events.
  • Broker submissions: 412 recorded attempts.
  • Recovery work: No separate reconciliation events occurred during this window, so reconciliation did not affect the average.

What the Standalone timing includes

The timer starts after Copilink admits a follower order event for processing and covers its internal follower-order work. It does not include broker, network, exchange, or internal queue-wait time. The count-weighted average across the observed events was 4.91 ms.

Follower gate and queue wait averaged 57.66 ms during this sample. That is scheduling contention before processing begins, rather than the processing work itself, so it is deliberately reported outside the ~5 ms processing average.

Copilink for NinjaTrader 8 benchmark

The published NinjaTrader 8 benchmark was conducted on a standard desktop PC to represent a typical user environment:

  • CPU: AMD Ryzen 7 2700X (Standard 8-Core Processor)
  • RAM: 16GB DDR4
  • Load: Five concurrent follower accounts receiving simultaneous copies.
  • Sample size: Performance data captured across about 70 live and simulated trades.

What the NinjaTrader 8 timing includes

Latency is measured at the application layer by timing the internal processing loop within the NinjaTrader 8 environment:

  1. Event Reception: The moment Copilink receives the OnOrderUpdate or OnExecutionUpdate event from the Leader account order.
  2. Command Submission: The exact moment the Copilink engine submits the corresponding orders to all 5 follower accounts.

The measured "Total Processing Time" covers the logic to identify the trade, calculate contract ratios based on follower settings, and submit the orders to the local NinjaTrader core.

Results

Across the NinjaTrader 8 testing phase of about 70 trades, Copilink recorded ~1.6 ms average internal latency. This represents the time Copilink takes to process and act after a leader event is captured in the measured flow.

Why This Matters

A clearly scoped internal benchmark helps traders evaluate the responsiveness of the copy engine without confusing it with factors outside the application. For Standalone, optional cloud reporting does not sit in the live copying path. In every case, the broker and market remain the source of truth for order acceptance, position state, and fills.

The fastest trade copier page puts both results next to how local and cloud copiers handle an order.