← Back to article listing

September 1, 2023
Share this

High-frequency trading (HFT) is a form of algorithmic trading where financial instruments, like stocks, index futures, are bought and sold at extremely high speeds.

Here are some key characteristics and components of HFT: 


Speed: HFT systems can make thousands or even millions of trades in a second. The trading decisions are made by algorithms, which can analyse market data, identify trading opportunities, and execute trades in fractions of a second. 


Short holding periods: Positions in HFT are typically held for very short durations, sometimes just a few seconds or even milliseconds. The goal is to benefit from tiny price inefficiencies that exist for very short periods. 


Co-location: To reduce latency (i.e., delays in trade execution), many high-frequency trading firms place their systems in the same data centres where exchanges host their trading systems. This practice, known as co-location, ensures that HFT firms can receive and act on market data as quickly as possible. 


Market data: HFT strategies often rely on detailed market data feeds that provide more information than what's available to the general public. For instance, some might use direct data feeds from exchanges rather than normalised consolidated feeds to get faster access to market movements. 


Strategies: High-frequency trading encompasses a variety of strategies. Some common ones include market making, statistical arbitrage, and trend following. However, there are also more controversial strategies like spoofing, layering and front running – these being illegal banned practices. 


Technology and infrastructure: High-frequency traders invest heavily in state-of-the-art technology. Every nanosecond counts, so having the fastest hardware, optimized software, and reliable networks is crucial. 


Controversy: HFT has been a controversial topic in the finance world. Advocates argue that HFT provides liquidity to the markets, tightens bid-ask spreads, and makes trading more efficient. Critics, however, contend that HFT can create market instability, disadvantage retail investors by front-running their orders, and lead to "flash crashes" where markets plummet and recover rapidly for no apparent fundamental reason. 


One famous incident often linked to HFT is the May 6, 2010, "Flash Crash" in the U.S. stock market. During this event, the Dow Jones Industrial Average plunged about 1000 points (around 9%) and recovered those losses within minutes. Though multiple factors contributed to the crash, HFT was identified as a contributing factor due to its rapid trading and the interplay of various algorithms. 


Since then, regulators in many countries have implemented rules, oversight and circuit breaker mechanisms to prevent market abuses and extreme events and ensure that HFT practices do not unduly harm market stability. 


The technology side of HFT is complex and highly competitive. Who wins the target trade is based on having the best trading algorithm strategy design along with the associated access to the low latency full market depth market data feeds and order entry APIs of the target liquidity pool(s). 


The FIX Protocol provides a degree of standardisation for these APIs, but low latency API access tends to be based on low latency binary level non-FIX protocols for speed and bandwidth efficiency. These APIs are generally unique to the venue and subject to ongoing change based on technical requirements and regulatory updates. 


This means that software development of trading strategies needs to support these APIs and to maintain them based on the constant venue protocol updates. 


The investment of time and money in development and supporting the direct market access (DMA) APIs is significant. 


This is where the OnixS offerings save time and money. 


The OnixS directConnect venue specific market data handler and order entry hander SDKs are ultra-low latency SDKs designed to be integrated into trading application frameworks. This offers efficiency in speed to market in developing and deploying trading frameworks development and reduces the development burden in the ongoing support based on using specialist DMA software development kits (SDKs).