Building the Artificial Trader

The demise of the cash equities trader has been reported for a number of years and many will claim that the excessive elbow room on the floor of the New York Stock Exchange (NYSE) proves the point. By Rob Daly

However, that is not actually the case. The trader's role has matured in lockstep with the advances in electronic trading. Those traders who performed the simple task of keying tickets into their firms' order management systems (OMSes) were the first victims to fall to automation. Soon after were the traders whose job it was to decide where to route their order flow.

Ever since the US Securities and Exchange Commission (SEC) introduced alternative trading systems (ATSes) and modified the order handling rules over a decade ago, traders have watched smart order routers take over another part of the trader's skill set.

Today, successful sales traders have gone from executing client trades to advising their clients on how to execute trades in order to meet the client's objective. In short, they have become execution consultants.

Their skill set now is knowing which of the variety of electronic execution strategies the broker-dealer offers will provide the client with best execution.

Once the client makes the decision, the order is usually sent on to the broker's smart order router for execution.

As trader's role has changed, so has smart order routing.

Initially, smart order routers were given the relatively simple task of sourcing liquidity within a limited number of displayed liquidity venues. Dealing with non-displayed liquidity, multiple clearinghouse relationships and amalgamated retail and institutional order flow weren't a concern at the time.

As these issues popped up, smart order routers have become more intelligent than ever before.

The need to have the smartest router on the Street will divide the sell side into those institutions that have the resources to develop their platform internally and those that will need to rely on a third-party vendor.

The best performance will likely come from those internally developed routers, where firms know every line of code under their router's hood and can make the necessary tweaks and changes to it without having to worry about a vendor's product release cycle or patch updates.

This isn't saying that third-party routers should be pitched, but firms need to realize that they will be using the same platform and order routing logic that a number of their competitors use.

The key aspect of all smart order routers is the ability to make custom changes to the platform. Most buy-side clients currently do not want to dictate their brokers' order routing strategies. It is not uncommon, however, for them to ask the brokers not to route their orders to one or two specific venues due to a variety of reasons.

A broker's router will need to be flexible enough to handle these one-off requests from their clients.

This may sound dire but it's not. If the third-party router vendors can deliver this capability along with a level of customization that would make each client's smart order router truly unique, they may give the internally developed routers a run for their money.

Rob Daly is editor of Dealing with Technology and can be reached at

For a free four-issue trial   of DWT, please visit: www.dealingwithtechnology.com -->