Pull up EURUSD, USDJPY and EURJPY on three adjacent charts. When EURUSD jumps, EURJPY should jump with it, almost simultaneously, since it's derived from the other two.
On a good feed, the lag is a few milliseconds and invisible. On a poor feed, EURJPY may follow half a second later, or in a few chunks. That delay reveals how quickly the broker's system updates its synthetic prices.
A script that logs the timestamps of ticks in each of the three pairs and calculates the delay between a move in the base pair and the cross is simple to write.
Why does it matter? A slow cross means the broker's price engine is under strain or built on cheaper infrastructure. The same weakness affects execution, where every order passes through the same system.
It also offers a way to compare brokers objectively. Running the same test on two feeds simultaneously gives numbers instead of impressions.
Some brokers show surprisingly large delays at busy times, with crosses lagging by a full second. The visible effect is a cross price that "catches up" in jumps.
Has a delay between the pairs and their cross been measured on a live feed?
How a stale cross exposes a laggy broker feed
-
Shadow Trader
- Posts: 5438
- Joined: Wed Sep 09, 2026 5:19 pm
-
Amanhasnoname
- Posts: 1506
- Joined: Fri Oct 30, 2026 9:45 am
Re: How a stale cross exposes a laggy broker feed
I tried this exact test on two brokers at the same time last spring. I wrote a small script that logged the time of each tick on EURUSD, USDJPY and EURJPY and measured how long the cross took to follow a jump in the euro. On the better broker the lag was usually too small to notice, a few milliseconds at most.
The other broker was a different story at busy moments. During the London open the cross sometimes trailed by a few hundred milliseconds and then caught up in two or three steps. At the US data releases I saw delays close to a full second, which lined up nicely with the poor fills I had been getting there.
It was the most convincing evidence I had ever collected for choosing between brokers, because it replaced a vague feeling with actual numbers. I moved my scalping to the faster feed and kept the other account only for slower trades where a second doesn't matter.
The other broker was a different story at busy moments. During the London open the cross sometimes trailed by a few hundred milliseconds and then caught up in two or three steps. At the US data releases I saw delays close to a full second, which lined up nicely with the poor fills I had been getting there.
It was the most convincing evidence I had ever collected for choosing between brokers, because it replaced a vague feeling with actual numbers. I moved my scalping to the faster feed and kept the other account only for slower trades where a second doesn't matter.