January 2014: Do Algorithms Dream of Super-Fast Sheep?

While algorithms are far from sentient, Victor wonders what they would think about if they could.

Apologies to those readers who scoff at my somewhat obvious play on the title of Philip K. Dick’s Do Androids Dream of Electric Sheep?, published in 1968 and which served as the premise for one of the finest science-fiction movies ever made: Ridley Scott’s Blade Runner. I’m not especially a fan of sci-fi movies as a genre, but few works come as close to cinematic perfection as does Blade Runner, which broods from one low-lit, rainy scene to another, set against Vangelis’ sublime soundtrack, and culminating in Rutger Hauer’s partially ad-libbed and now iconic “Tears in the Rain” monologue. Yes, I’m a big fan.

But there’s more to my rhetorical question than simply conveying my appreciation of Blade Runner. Since their introduction to the capital markets more than a decade ago, algorithms have been evolving in terms of their sophistication, from basic computer programs designed to fire buy and sell orders into the marketplace when the optimum conditions occur, to entities that possess a certain artificial intelligence that allows them to undertake far more than merely financial firms’ donkey work.

Now, algorithms have the ability to “learn” from past “experience,” intentionally routing orders to specific destinations to obtain the most advantageous and likely fills. While this behavior is far from anything approaching a consciousness and the ability to genuinely make their own decisions based on something other than programmed inputs, surely that time will come sooner or later.

In the meantime, the capital markets remain awash with hundreds of thousands of algorithms watching and waiting for their opportunities to strike. While that characterization might sound a little sinister—and to those commentators who point the finger of blame at rogue algorithms responsible for fiascos such as the Flash Crash, I’m sure it does—it’s important to remember that as intelligent as they are, they are really only the coded manifestations of quants’ calculations, designed to carry out their work in a logical and consistent manner. The fact that algorithms’ behavior is closely tied to many of the most serious technology glitches to have occurred in the last five years is irrefutable. But if they could defend themselves, they’d be quick to argue that even the rogues among them are doing only what their human masters have commanded, and no more.

And so, while we might still be some way off from witnessing the introduction of genuinely intelligent algorithms that possess an artificial consciousness, it doesn’t hurt to speculate about what they might end up dreaming about in the early hours when their minders have left the office.

• Tweet  
• Facebook  
• LinkedIn  
• Save this article
• Send to  
• Print this page  

• Topics
• Algorithmic trading
• Trading Tools
• Management & Strategy
• Flash crash