Automated trading systems

Finance - Automated Trading Systems

AUTOMATED TRADING SYSTEMS

The Mathematical Basis of Securities Trading:

In principle, the concept of securities trading is readily reducible to mathematical relationships capable of predicting future stock performance on the basis of previous trends in the market (Skypala, 2006). However, myriad factors that do not immediately appear relevant to the stock market also have great potential to influence the market.

This has always been true, (as illustrated very simply by the relationship between wars or other serious global occurrences during peacetime throughout the 20th century), but since the computer age and the 24-hour news cycle, events external to the market play a much larger role than before (Duhigg, 2006). Nowadays, "speed is everything" according to a solutions executive and financial market analyst at IBM (Skypala, 2006). One of the primary determinants of the maximum speed of trade execution are the algorithmic trading systems and the relative speed with which they permit brokers to receive and analyze market data and then execute the trade. Processing speed is a crucial component of this process (Skypala, 2006), because the current trading climate is not unlike an infinitely more complex but otherwise similar dynamic to that experienced by E-bay bidders attempting to outbid competitors by the smallest margin of time possible to eliminate subsequent bidding before the close of auction.

For decades, Wall Street firms and hedge funds have recruited analysts with advanced degrees in mathematics, engineering, and physics because their skill sets and natural aptitudes are conducive to identifying highly complex market patterns capable of being exploited for profit. Specifically, the algorithms devised by these analysts enables them to determine subtle relationships between many potential factors that present opportunities for favorable trades. In many instances, these types of analyses compare similarly situated companies, such as Microsoft and IBM, and suggest potential profits by selling one while buying the other based on their small term divergence in their stock prices in conjunction with expectations (based on historical patterns of synchronicity between the two stocks) that those temporary fluctuations will correct themselves subsequently (Duhigg, 2006).

The combination of high-speed computer processing with complex algorithms led many to predict the eventual extinction of the (human) stock broker as soon as computers were first introduced to the securities industry more than two decades ago. However, even those observers failed to anticipate the degree to which processing speed and worldwide interconnectedness of thousands of computers would lead to the evolution of a largely automated 24-hour news cycle in which computers search for and identify relevant news information without human help (Dey, 2006).

This tremendous increase in both the volume of factors potentially relevant to the securities market and its automated nearly instantaneous transmission, reception, and analysis have dramatically increased the complexity of financial markets. In many respects, the situation that evolved has begun to resemble a computerized "arms race."

According to Andrew Lo, Director of the Massachusetts Institute of Technology's Laboratory for Financial Engineering, "Everyone is building more sophisticated algorithms, and the more competition exists, the smaller the profits" (Duhigg, 2006).

The Evolution of Automation:

In the first decade of the 21st century, the increased processing power of computers and the field of artificial intelligence are already poised to revolutionize the securities industry, possibly even excluding humans from most aspects of trading within a very few years. Already, computers track billions of market transactions and identify complex patterns and relationships that defy human intellect (Duhigg, 2006). According to Louis Morgan, Managing Director of HG Trading in Wisconsin, "Five years ago it would have taken $500,000 and 12 people to do what today takes a few computers and co-workers... I'm executing 1,500 to 2,000 trades a day and monitoring 1,500 pairs of stocks. My software can automatically execute a trade within 20 milliseconds - five times faster than it would take for my finger to hit the buy button" Furthermore, whereas many of the successes of current generations of market analysts have emphasized their mathematical talents, the latest technological advances in the field include sophisticated software such as the Apama Algorithmic Trading Platform that now enables reasonably skilled day traders to construct their own complicated trading algorithms "almost as easily as they drag an icon across a digital desktop" in fact, studies suggest that as early as 2005, as many as one-third of all stock trades in the United States were being initiated by automatic computerized algorithms and that the shift to automatic processes was responsible for the "explosion in stock market activity" between 1995 and 2005, when the average daily volume of shares traded on the New York Stock Exchange increased to 1.6 billion from 346 million. (Duhigg, 2006).

In particular, the union of computational capacity and the instantaneous, automatic availability of worldwide news in real time helped pave the way for the most recent surge of financially profitable securities trading strategies. Consider that analysts can now program their systems to turn around a report on a company's financial results within 0.3 seconds of that information being issued to the Stock Exchange (Dey, 2006).

However, what actually accelerated the increased reliance on computer-initiated trades was the slightly delayed evolution of methods of hiding large trades from other computers rather than their lack of ability to analyze the market and initiate responsive trading automatically (Temperton, 2007).

It was the more recent introduction of so-called black boxes that enabled traders to build major stakes in corporations or move conspicuously large orders without drawing attention from competitors. Similar large-scale stock movements via traditional methods (whether by phone or even automatically through very fast, sophisticated computers) run a significant risk of pushing the price of the target stock down by virtue of the mechanism called "slippage" within the industry (Dey, 2006).

By dividing up the transaction into smaller movements, this more stealthy approach has allowed automatic computer trading to maximize the fullest potential of computer processing power without jeopardizing the profit margin by alerting the market to large trades (Curran, 2008). Likewise, these mechanisms allow purchasers to accumulate large stakes in companies without unnecessarily disclosing their interests until they reach the statutory limit of three percent before disclosure is required by law (Dey, 2006). One algorithm introduced by Credit Suisse called "Guerrilla" is designed to identify in real time any publicly displayed bids or offers on an exchange or trading platform that are ripe for exploitation with a minimal risk of triggering jumps or displacements in the stock's established trading patterns. In essence, this enables fund managers to avoid "moving prices against themselves" (Temperton, 2007). Naturally, however, as such techniques became more common in recent years, their yields have slowed down (Duhigg, 2006).

Benefits Associated with Automated Trading Processes:

According to industry analysts, the "booming trading figures" recently witnessed in the London Stock Exchange are partly attributable to so-called black box technology.

Because the London Stock Exchange charges commissions on a price per-trade basis, the automated division of large trades into many smaller trades triggers increased commissions without the requirement of corresponding trading volume increases. Beyond its larger beneficial effects on stock markets, some observers believe that the continuing shift to execution of "bread-and-butter" trades has enabled investment banks and brokerages to cut labor costs in addition to freeing up analysts to devote less time to trading and more time to identifying potential opportunities and developing new investment strategies (Dey, 2006).

Temperton (2007) describes algorithmic techniques of automatically ensuring the capture of a specific proportion of the trading volume of targeted stocks that are particularly useful to" momentum-based" investors and fund managers who focus on trends in volume as an indicator that often corroborates price trends. Another technique is applicable to indexed-based or enhanced index-based strategies: "benchmark" algorithms allow the establishment of predetermined benchmark predicated, for example, on short latencies, such as in relation to the volume weighted average price over a certain time period. In conjunction with "iceberging" strategies for hiding large stock movements, these capabilities represent powerful profit-making tools for investors (Temperton, 2007).

Risks Associated with Automated Trading Processes:

Analysts caution that greater reliance on automated processes also presents opportunities for costly mistakes. According to Stacy Williams, Director of Quantitative Strategies at HSBC Global Markets, "The downside with these systems is their black box-ness...traders have intuitive senses of how the world works. But with these systems you pour in a bunch of numbers, and something comes out the other end, and it's not always intuitive or clear why the black box latched onto certain data or relationships."

Williams, who is involved in predictive modeling studies at Cambridge University detailing the way computer viruses relate to foreign currency market forecasts at also implies that as automated trading becomes more common, the risks associated with it increase correspondingly: "Only an elite group of people are using these ideas, but a lot of people are thinking about them" (Duhigg, 2006).

Similarly, one industry analyst suggests that "greater reliance on sophisticated technology and modeling brings with it a greater risk that systems failure can result in business interruption" while another characterizes the comprehensive relative risks and benefits thusly:

The fund management firms don't need to staff up to the same extent. The algorithm can be your market eyes. it's effectively a trading assistant - a very diligent trading assistant... The downside is that it is also a very obedient trading assistant, so if you tell it to do something it might not have the intuition or the ability to veto you... obviously there are checks and balances to prevent anything bad from happening, but you do hear stories about people putting an order in with the wrong instruction, it moved the stock 10 per cent and then you get a call from the regulator" (Dey, 2006) in 2007, the Economist attributed a financially significant "wobble" suffered by the New York Stock Exchange on February 27, 2007 to the ad hoc combination of increasing capacity by adding more scalable hardware to a system that still relies substantially on floor-based trading, yielding a "hybrid" system with significant vulnerabilities. According to that journal, the anomaly produced in the NYSE system on February 27th, NASDAQ absorbed the shift, slowing down NASDAQ as well, simultaneously illustrating certain volume-based vulnerabilities in that system suggesting the need for upgrades to keep up with technological capabilities of automated trading (the Economist, 2007). Implications for the Future: