Who Should You Blame for High Latency? When It’s Your Broker’s Fault and When It’s Not…

Trading happens at near light speeds – low latency trading is critical to engage the marketplace with maximum efficiency.

Abrupt and swift pricing fluctuations can happen in microseconds. To compete in this hypercompetitive arena, a trader must reduce as much trade-related latency as possible.




What Happens When There is High Latency?

Simply put, trade-related latency is the time it takes for an order to be executed, from a signal sent, to a fill at the exchange. The trade is successfully executed at the clicked price assuming there is no latency in the network when the order is routed back to your broker’s server.

If the order message that is routed back to your broker’s server gets delayed even for a millisecond, the price you attempted to complete the trade at may have moved away and the broker will not be able to fill the order.

The same latency can occur with API or high frequency trading.

On top of delays in order execution, high latency can also have a serious effect on price updates resulting in traders seeing “delayed” or “stale” prices, no longer reflecting current market data – leading to slippage.


When Latency Is Your Broker’s Fault – And When It’s Not…

Latency is usually caused by the distance between communication equipment and network congestion.

It can be found in brokerage servers, internet connectivity, software, hardware, and broker’s network infrastructure.

Let’s compare a trading network to a busy highway…

  • Bandwidth / Internet Connectivity: determines how narrow or wide the information highway is. The narrower it is, the less data it can handle at once, causing network congestion.
  • Latency: measures how fast the information travel from Point A to Point B and back within the highway.
  • Throughput: is the amount of traffic that can be handled by the highway over a given period.

If one of these highways get bottlenecked, a “data lag” can occur and this data lag can mean the difference in favorable fill prices and slippage. Your first step should be to look inward for the lag by running a speed check on your end, but you are also entitled to ask your broker for a “speed-check.” They have reliable tools that can give you an honest reading and potentially tell you if they are causing some or all your latency issues.




How Do You Achieve Low Latency?

Traders will always be in an uphill battle when trying to outrun the big institutions. The institutional players invest huge amounts of capital into latency-reducing tech, and their resources are not typically available to traders. When it comes to speed, the playing field is not perfectly level. However, there are several steps traders can take to help reduce trade-related latency that will help improve the bottom-line while increasing the overall chances of profitability.


Here are a few steps:

  • It’s not you, it’s me: Check and maintain your hardware. Updates and maintenance should be part of your routine before you go looking for any other problems with high latency.
  • Your connection to the exchange: Perform “ping” tests and run a “traceroute” to the appropriate servers to check the stability and robustness of your connection.
  • Server-side hosting: Server-side hosting helps to reduce outages and eliminate any client-side communication issues, thereby reducing latency.
  • Automate, automate, automate: The closer your strategy gets to being fully automated, the less time it will take to hit the market, and you can then potentially use collocated servers at the exchange. Closer means faster.
  • Direct market access (DMA): Orders are placed directly at the exchange without first being routed through an intermediary. Anytime you reduce the number of steps, you increase speed, making DMA an obvious solution to help reduce latency.
  • VPN: VPN can help reduce latency, as shown in the diagram titled “Lime’s Low Latency Measurements vs a Retail Competitor.”

In “Preferred Low Latency Setup Diagram,” you can see the normal flow of a retail trader’s order. This diagram shows a trader who has a collocated server, and another shows the flow for the average trader that sits anywhere in the world. Within this flow several of the culprits of speed are described above.


Preferred Low Latency Setup Diagram

Clearly having a collocated server takes care of some of the in-house problems (there is still an issue with location of the server), so let’s remove the two clients and look at the flow of the order where these two fictional clients end up being equal. This is the section of “Example Diagram of Low Latency Connections into Lime” where your broker could be responsible for your latency problems. As shown here, sometimes the broker is not collocated at an exchange. If you are having latency problems, this is clearly a time when it’s the “broker’s fault.”


Example Diagram of Low Latency Connections into Lime

In “Lime’s Low Latency Measurements vs a Retail Competitor,” you can see a breakdown of how long some of these flows take. It’s plain to see that a collocated server like the one provided by Lime can help reduce latency on the broker’s end. Even a VPN into an agency broker like line is much faster than your typical retail order flow.


Lime’s Low Latency Measurements vs a Retail Competitor

When dealing with high latency, discovering the problem is 90% of the battle and oftentimes the answer is location, location, location.