Algorithmic Trading: Risks and Realities of Smart Investing

What is algorithmic trading?

Algorithmic trading is exactly what it sounds like. Instead of making trading decisions based primarily on human knowledge and intuition, a trader or investor offloads the process onto computer algorithms. These algorithms are essentially a set of predefined rules and processes that instruct a trading program to buy or sell based on certain triggers or conditions.

These triggers might include factors such as elapsed time, price, quantity and trading volume. Most automated trading algorithms also align with a particular trading strategy, such as trend following, arbitrage or index fund rebalancing.

Taking the stock market by storm

Algorithmic trading is not a new practice. Although today's stock trading algorithms are considerably more sophisticated, the technology has been around since at least the early 1980s. It's also worth noting that the stock market operates via its own sophisticated algorithms and has done so for decades.

It's therefore no surprise that algorithmic trading has gained so much popularity. To some extent, investment and trading has always been a data-driven sector. Algorithmic trading simply represents a new iteration of the industry, one which addresses the greatest weakness in legacy stock trading: people.

No matter how skilled and experienced a stockbroker may be, they are still fallible. They can misinterpret data and execute on non-profitable, non-viable assets. And even someone who's otherwise knowledgeable might make decisions based on emotion and intuition rather than logic.

Automated trading software does not possess intuition. It lacks the capacity for emotion or independent thought. All it knows how to do is follow its own predefined instructions.

Accuracy aside, there's also the matter of efficiency. Algorithmic trading allows investors and funds to operate at a scope and speed beyond any human being. It operates in real-time, executing trades at the best possible moment and for the best possible price.

This efficiency consequently allows investors to more effectively profit from market volatility while also ensuring greater liquidity.

Tapping into a volatile market

There are several ways an investor might leverage trading algorithms to take advantage of market volatility.

High-frequency trading

High-frequency trading represents one of the most common forms of algorithmic trading, identifying patterns and executing trades in real-time. A fund leveraging high-frequency trading takes advantage of millisecond-by-millisecond price movements, executing trades as those movements occur. In so doing, the fund profits from price changes that other firms either failed to notice or lacked the capacity to exploit.

Statistical arbitrage

Statistical arbitrage, or stat arb, is one of the more sophisticated algorithmic trading strategies. A fund that leverages stat arb typically maintains a massive securities portfolio that it trades over an extremely short timeframe, typically either a few seconds or a few days. Through market analysis and sophisticated computer models, the fund also assigns a desirability score to each stock, then groups assets together based on this score.

Mathematical and computer models may also be applied to any of the other trading strategies, as well.

Algorithmic options trading

An algorithmic options trading strategy consists of two parts. First, the fund builds a portfolio of options, opting to hold on to each rather than selling. Then, through sophisticated trading algorithms, the firm predicts changes in market volatility, using that foreknowledge to determine when to buy or sell.

This strategy typically pairs very well with a Delta neutral portfolio strategy, which involves the creation of a portfolio where negative and positive value changes cancel one another out.

Mean reversion

Also known as a trading range strategy, mean reversion is predicated on the belief that price fluctuations in an asset are temporary and that every asset will ultimately return to its baseline price. A fund that employs a mean reversion strategy must first define the price range for each asset.

Once this is done, the fund can then deploy an algorithm that automatically executes trades on an asset when it drifts from its defined range.

Weighted average price

A weighted average price strategy combines historical data with algorithmic trading to dynamically release larger orders in bits and pieces. In the case of a volume-weighted strategy, the end goal is to get as close as possible to the average value of an asset over the course of day. A time-weighted strategy, meanwhile, seeks to minimize market impact by releasing a security at regular intervals.

Derailing the hype train

To say that the world is currently obsessed with artificial intelligence and machine learning is putting it lightly. People are either excited about the technology beyond all rhyme and reason, or terrified of its potential implications for the economy and the job market. As is often the case, this degree of hype gives itself over quite well to misinformation — such as the assumption that algorithmic trading has its roots in AI.

Artificial intelligence is effectively an evolution of the computer algorithm. Even the most sophisticated AI at its core is a collection of complex, dynamic algorithms which together can adapt and change based on external data. To put it another way, algorithms define a process or set of actions based on a specific trigger, while artificial intelligence determines how to deal with that trigger on its own.

While there are segments of the algorithmic trading space that veer into the domain of AI and ML, algorithmic trading is still distinct from AI. It's also subject to its own unique brand of hype, one which obscures the technology's shortcomings.

Black swan events

While historical data allows trading algorithms to accurately determine future market movements in a vacuum, it's ultimately impossible to actively predict every change. This means that a trader who is over-reliant on algorithmic trading could suffer crippling losses at the hands of a truly unforeseen disruption, known as a black swan event. Such events have several primary characteristics.

First, black swan events are extremely rare and impossible to predict, existing as complete outliers. They are also highly disruptive, often causing catastrophic damage to multiple markets and economies. Finally, when examined in hindsight, the circumstances that created a black swan event seems incredibly obvious — so much so that it's difficult to believe no one saw it coming.

The best way to protect yourself against black swan events is to understand that for all it has to offer, algorithmic trading is still fallible. You need to plan for the worst and ask yourself what happens if you encounter a market disruption that you fail to predict.

Negative market impact

Used unscrupulously or carelessly, algorithmic trading solutions can greatly increase market volatility. Not only does this make things increasingly difficult for traders that don't leverage the technology, there have also been circumstances where algorithmic trading has been suspected of causing flash crashes, which are essentially self-contained market disruptions that typically occur within a single trading day.

Make sure that whatever predictions you make also include the potential impact you might have on the market.

Complex regulatory oversight

Regulatory compliance has been the source of many a headache for investors and business owners alike.

Unfortunately, depending on where you're situated, you might want to reach for the aspirin. In the United States, for instance, firms that engage in algorithmic trading must adhere to both SEC and FINRA regulations. Algorithmic trading in the EU, meanwhile, is covered under MiFID II and MiFIR.

When it comes to contending with regulatory compliance, you have two options. Your first option is to familiarize yourself with the regulations on your own for each market in which you intend to operate. Alternatively, you could simply hire a compliance officer to help yourself stay on track.

Over-reliance on computing infrastructure

Algorithmic trading relies on a complex web of technology. If anythingin that web ceases operations, everything comes screeching to a halt. As you've probably guessed, this could result in significant losses for an investor.

Redundancy is your best bet. Identify any single points of failure in your infrastructure and ensure that you have backups available in the event that something stops working. Develop and regularly test an incident response plan, as well. Finally, diversify your investments — do not simply operate a portfolio filled with volatile, short-term assets.

Lack of human intelligence

Algorithmic trading is a powerful tool, but it's not something your organization can simply leave to run unsupervised. The majority of algorithmic trading tools are built on a series of predefined rules, which means they fall somewhat short in terms of flexibility. That aside, even the most sophisticated algorithm still requires occasional human input and oversight.

Algorithmic trading should never fully replace human intuition and judgment. While there's certainly nothing wrong with relying on algorithms for market analysis and workflow automation, you must also be prepared to occasionally step in with a guiding hand.

A landscape built on intelligent investing

The majority of firms and funds are at least familiar with algorithmic trading. However, as is often the case, there are a few organizations that stand out. Some of them are well established, while others are relatively new to the market.

However, they all share one common characteristic — an innovative mindset.

Areus Asset Management, for instance, has taken a unique approach to investing in the spot gold/USD currency pair. By leaning on sophisticated algorithms for efficiency, accuracy and risk reduction, it has constructed a promising absolute returns portfolio. The company currently aims for double-digit absolute growth and it's easy to imagine them achieving that goal.

Since its launch in March 2022, Areus Asset Management has already reached a significant milestone in its first 12 months of trading, with its underlying portfolio returning more than 48.91 percent to investors between March 1, 2022, and February 28, 2023.

63 Moons Technologies (NSE:63MOONS) is arguably the biggest software vendor in the algorithmic trading market, courtesy of its ODIN Algorithmic Trading software. Armed with both powerful, purpose-built algorithms and built-in artificial intelligence, the software is capable of performing complex calculations at microsecond speeds.

As far as high-frequency trading is concerned, Virtu Financial (NASDAQ:VIRT) remains the reigning titan, particularly in the wake of high-profile acquisitions such as KCG Holdings and ITG. The firm prides itself on combining its market making expertise with highly sophisticated technology. Services offered by Virtu Financial include analytics, liquidity sourcing, execution, analytics and broker-neutral, multi-dealer platforms.

As the market making arm of hedge fund Citadel, Citadel Securities combines comprehensive expertise with advanced predictive models and analytics technology. Working with some of the world's top financial institutions, the firm manages over US$30 billion in assets worldwide and executes US$458 billion in trades each day. It also prides itself on attracting some of the world's best tech talent, with nearly half of all employees holding an advanced degree.

Takeaway

Algorithmic trading is a good investment strategy. However, like any strategy, it isn't perfect. While it has the potential to generate immense value for a savvy investor, attempting to leverage the technology without understanding it is a recipe for disaster.

With that said, current trends appear to indicate that this technology might ultimately be the future of the stock market — as such, investors and financial services professionals alike should start building knowledge and skills sooner rather than later.