The end of high frequency trading, the rise of algorithms


On Wall Street, a pet peeve of more than a few executives is the tendency, even in the financial services industry, to conflate high frequency trading and algorithmic trading. To be sure, there are algorithms that depend on raw speed. And some high frequency traders, perhaps those who seek rebates almost exclusively, have automated some processes.

But for the most part, the two approaches to mining the market for profits are distinct.

Quantopian CEO John "Fawce" Fawcett has some interesting thoughts on this.

"The speed wars are over. Finance is moving from the high-frequency era to the big-data era," Fawcett notes. "The capital costs for high-frequency trading get higher with each millisecond, and the rewards get smaller. And the regulators are coming – high-frequency trading isn't going to be the same when they are done."

So what's next?

"High-frequency trading has been hogging the spotlight, and most people haven't noticed the other, bigger benefits of investing with algorithms. Algorithms bring reproducible backtests, letting traders test against historical data, giving the opportunity to weed out many losing strategies before they commit capital.

"Algorithms remove human emotion and bias from trading decisions, while marrying human insight and strategic thinking with speed, breadth and rigor. The literature is filled with the mistakes that humans make when they get attached to a position or a company or a trade; an algorithm never falls prey to these cognitive biases.

"Algorithms permit a gigantic increase in the scale and scope of trading decisions. With an algorithm, you can evaluate thousands of securities using thousands of distinct data sources, something that is impossible for humans to do by themselves. The future of finance is algorithms looking beyond the order book to the rapidly expanding universe of time-series data."