Over at Human Benchmark, a reaction test finds that the median reaction time for those taking a simple test is about 215 milliseconds (much better than my score of 280 milliseconds, but then I like to take my time). That’s pretty darn fast — just a little longer than one-fifth of a second. But when it comes to financial transactions, it’s an eternity.
You may have seen footage of stock exchanges where traders hold paper up in the air and scream “buy” or “sell” or “did someone drop this paper?” But these days, the big companies aren’t using humans to shift stocks in companies. Instead, they’re using computers and special algorithms.
An algorithm is just a set of instructions. It gives a computer the direction it needs to perform operations. In this case, the algorithms are designed to respond to changes in the stock market — even tiny changes. And they do it at a speed that’s mind-blowing. We’re talking less than a millisecond for some transactions.
Think about that. These algorithms look for subtle shifts in the market and then can pounce on an opportunity before we humans even know something is up. And if enough different computers act on this, we can get a little miniature stock boom — or crash — and it can all happen in less than two seconds.
Researchers have been studying how these ultra-fast machines change market conditions and the results are interesting (or terrifying, depending upon your point of view). They found that between January 2006 and February 2011 more than 18,000 “extreme events” occurred within the span of 1.5 seconds. According to the report, these events were caused by algorithmic trading and included both spikes and crashes in stock price.
Besides creating great depressions on the millisecond scale, the ultra-fast trading may also contribute to some computer problems the global markets have been having in 2013. It’s possible that when thousands of computer programs all begin this type of trading it can overwhelm the computer systems in charge of tracking the market as a whole, crashing them in the process.
I’m sure the fans of science fiction are way ahead of me on this — but imagine a scenario where these algorithms fixate on a particular stock as it experiences a fluctuation in price. The robotic traders pounce at a blinding speed and the tracking computer suddenly sees a tiny fluctuation become a dramatic change. Soon, related stocks have similar fluctuations as traders — both human and robotic — begin to worry about a shift in an entire industry. Next thing you know, your 401(k) gets annihilated by Robby The Robot Day Trader.
I don’t think that’s going to happen, necessarily. But with recent events like the real estate bubble behind us, it might pay to look more closely at these algorithms and ensure they aren’t just a little too eager for their own good.