Commentary

Tim Quast
Traders Magazine Online News

We're All HFTs Now

In this guest commentary, author Tim Quast looks back at the history of HFT and how the market has evolved to where many firms now fit the definition of high-frequency trader.

Traders Poll

Are you in favor of a pilot program and examination of the rebate system by the SEC?




Free Site Registration

May 16, 2007

Morgan Stanley Seeks One-Upmanship with New Algo

By Nina Mehta

Traders can't yet be cloned, but Morgan Stanley is hoping to move a step in that direction with One, a customizable algorithm for clients that can be used to replicate how a particular trader actually trades a stock.

Morgan Stanley's One switches automatically from algorithms to direct market access and back again to meet a trader's specified goal. It can be used to execute preset strategies or strategies customized to an individual trader, said Andrew Silverman, the big broker's head of U.S. electronic trading distribution.

Morgan Stanley can develop customized strategies for clients essentially overnight. "A customer can tell us how he wants the algo to behave-exactly, down to the basis points-and we'll turn the strategy around in a day or two," Silverman said. The client can also decide which markets to access or not access, where to post orders and which dark pools to tap for liquidity.

One circumvents two problems that beset traders using algos, Silverman said. The first is that they must select the right algorithm for a strategy and input the right parameters to meet the execution demands made by portfolio managers. The second is that traders must baby-sit algos to make sure they're executing appropriately under changing market conditions.

The new algo gets around these issues, according to Silverman, by deploying a strategy tailored to the trader's precise needs in the first place. "It's a waste of a trader's time to notice a stock dropped 20 basis points, stop trading through an algorithm, pull the order out of the algo, trade it [manually] through DMA and then put it back into an algo," Silverman said. "We can take care of that for them and do it on a per-trader basis."

A trader, for example, may be comfortable buying a stock until its price has risen 20 basis points, but will then want to trade the stock by being 20 percent of the volume. That trader's One algo can do that. Another trader's One algorithm may follow an arrival price strategy with a high urgency level, but if the stock being bought drops 50 basis points, the algo will immediately grab as much stock as possible in both the displayed and dark markets.

One, under development for the last four months, has been tested internally by Morgan Stanley and by a large trading firm. It will be rolled out to clients in the coming weeks.