Despite the growing use of algorithmic trading, traders themselves still have great careers ahead of them, said several experts at the TradeTech USA 2006 conference in New York . But traders should get training in the use of algorithms if they want to avoid being sidelined.
While most buy-side traders have algorithms on their desks, only about five percent of orders use algorithmic engines, said Joseph Wald, chief executive of EdgeTrade, an independent agency-only trading firm that provides algorithms to traders on both the buy side and sell side. "The best algorithms do what computers do well and don't attempt to do what people do well," he said. "They can look at massive amount of data and respond in milliseconds. They also have a lot of patience - they can sit and watch prices all day long."
Although algorithms "cannot exercise judgment or respond to a scenario that hasn't been defined," they do complement a trader's skills, said Wald. He divides traders into two groups: those who have learned to use tools like algorithms, and those he sees as at risk because they don't understand the playing field is no longer level.
Chetan Joglekar, senior trader at the $370 billion pension fund TIAA-CREFF, said his trading desk builds its own algorithms but doesn't expect them to do everything correctly. In a basket of 500 stocks, a trader may have to step in and finish 50, he said. Unless a trader is exceptionally busy, it's rare to put a stock into an algorithmic engine for an entire day. And, Joglekar adds, a key trading skill is knowing when an algorithm won't work well. For example, if a company just reported bad earnings, an algorithm may run into trouble because it's based on historical earnings. Algorithms are also no substitute for traders when it comes to low-cap issues.
At Teacher Retirement System of Texas, Chief Equity Trader Claudia Williams says she examines each potential trade for risks, considers momentum, then sends potentially difficult trades to the order desk.
Greg Komansky, vice president at Legg Mason Capital Management, disputes the idea that algorithmic trading is like a device you can set and forget. "You put in the trade and then you need to watch the stock," he said. "I can't stand it when people say firms will have just two traders and a black box. I really don't agree. You need to watch and be mindful of every single trade."