On Monday July 6th 2009, various news outlets are reporting on the rather brazen bank theft by one Sergey Aleynikov. Rather than brandishing a gun or cracking a vault, Sergey hacked the algorithmic trading secrets of his then-employer Goldman Sachs by downloading proprietary, "black box" computer models that Goldman uses to execute rapid-fire trades in the financial markets. The value of this intellectual property, experts say, could be incalculable.

Toomre Capital Markets LLC ("TCM") has written extensively about the topic of Algorithmic Trading. Interested readers, for instance, might want to review the white paper entitled Market Risk and Algorithmic Trading that TCM wrote on behalf of Advanced Micro Devices some months ago. As that paper starts,

In the evolving financial markets, ever-more complex quantitative analyses are performed. Some constantly assess the market risk of portfolio exposures, while others calculate the probability of reward for various strategies in the continually shifting markets.

Increasingly, algorithmic trading programs automatically execute the trade orders that result. With the growing adoption of the AMD Opteron™ processor, high performance computing for quantitative modeling and algorithmic trading in the financial markets likely will increase.

Simulation modeling techniques quantify market risk, measuring the probability and magnitude of potential loss due to change in prices. As market liquidity decreases, typically price volatility and, hence, market risk increases. With the recent introduction of decimalization, the U.S. equity market structure dramatically changed. Trading spreads shrunk, trading venues proliferated, and market liquidity fractured. As a result, a new form of trade execution emerged: algorithmic trading.

Wall Street firms like Goldman Sachs and others have devoted considerable intellectual and technology capital to trying capture as much of the shrinking trading margin as they could. For instance, if any person or firm could create the technology infrastructure and analytic code so that one was able to react the fastest to any price event and actually consistently execute a buy or sell order as a result (before other market participants could do so), one theoretically could capture a significant portion of the pricing inefficiencies that occur in the normal course of the real-time trading of a security.

For the last several years, Goldman Sachs has been rumored to be quite good in executing algorithmic trading strategies, particularly for its own proprietary trading accounts. Hence, some large institutional investors have supposedly steered their "flow" execution business away from Goldman Sachs so that the "black box" algorithms did not use pending order information to potentially "profit" from the information that a certain amount of common stock might be bought or sold over the next X period. Goldman Sachs does not specifically break out where the profits from its equity division are derived. However, the profits from that division of the firm are considerable and a significant portion are said to result from algorithmic trading activities.

The net result is that the possible use of its proprietary computer code by Sergey Aleynikov at his new employer Teza Technologies (run by the ex-Citadel Head of High-Frequency Trading and reputed trading genius Misha Malyshev) potentially could compromise Goldman Sachs' future algorithmic trading profits significantly. Hence, that probably is why the Federal prosecutors came down so heavily on Mr. Aleynikov. One wonders though what part of the uploading of files to a foreign computer had to do with open-source software projects and what part was truly Goldman Sachs specific proprietary code.