CA Cheuvreux, Credit Agricole's European broker and a wholly-owned subsidiary of CA's brokerage Calyon that trades in 60 countries, has an ambitious global growth strategy in which the U.S. figures prominently. The firm provides DMA, program trading, algorithmic trading and CFDs, and other execution services to hedge funds. It sells electronic trading services and trading algorithms (sometimes white-labeling them) to U.S. broker/dealers. It also seeks to provide synthetic prime brokerage services to high-frequency trading firms.CA Cheuvreux recently hired Matthew Rowley, the designer of Fidessa's BlueBox algorithm toolbox, to be its new U.S. head of quantitative trading. We met up with Rowley in CA Cheuvreux's midtown office to find about more about his new role.

WS&T: What is your background?

Rowley: I went to Oxford University for mathematics, then I focused on statistics post-graduate. From there I went to First Chicago, and worked on derivatives in a small quant team. That was my first career move in the industry. Some of the mathematics for derivatives is fleshed out and quite commonplace now, but the techniques were still quite secret then. That was good exposure for me early on. First Chicago got taken over by Bank One, which eventually got taken over by JPMorgan. Bank One was more conservative at the time so they decided to reduce their derivatives risk. An opportunity at Fidessa came up (they only had 18 people then, now they have more than 1,300). I was at Fidessa eight and a half years and originated the project design for the company's BlueBox application, a set of development tools and algorithms. Cheuvreux, which was using BlueBox in the U.S. to achieve quick time to market, decided to bring in the guy that designed it.

WS&T: What does your new job entail?

Rowley: My mandate now is to help integrate the intellectual property CA Cheuvreux has in algorithmic trading in Europe and the U.S. CA Cheuvreux has been involved in algo trading now for seven years and has a team of 15 quants in Europe. I need to bring great ideas from the U.S. to the team. The idea of market fragmentation and dark pools is new in Europe, the U.S. is a few years ahead in terms of lit markets and smart routers. So the guys in Europe are leveraging off our knowledge. In September, CA Cheuvreux will launch a liquidity-seeking algorithm in Europe that I've been heavily involved in. We use sophisticated techniques to rebalance the orders out to the pools, depending on the fill rates coming back. We've seen an upswing in the use of these algo types. People still use VWAP and percentage and volume algorithms, but the use of those dark pool aggregation algos, smart algos that can seek liquidity, is in demand.

WS&T: What other innovations are you working on?

Rowley: Here's something really exciting we've been working on: using machine-readable, XML-tagged news feeds for better execution. There are two stages to this project. The first is to identify automatically very strange, highly volatile days, and react immediately, faster than traders could, to news items that will lead us to believe the stock will move quickly. The algorithm will then choose pre-set rules based on those market conditions. The second phase is to derive directional signals, whether the market is going up or down in the short term, and automatically change the parameters of the algorithm. A trader would take a second or so to process the information and another second to press the button and organize their blotter, while this will react in milliseconds. We want to allow clients to have fewer traders and more technology, or the same number of traders with more efficiency.

WS&T: Will customers pay more for this or do you think it will attract new customers?

Rowley: I think the quality of execution will improve, and if we can improve the quality of executions across a number of different market conditions, that will attract more customers. We also think that if we can prove that the algos work across a number of thinner name stocks that don't trade very much, then maybe we can attract more order flow that way.

WS&T: What other projects are you working on?

Rowley: We've built a market simulator with our Paris team as well as heat mapping technologies.

WS&T: What is the heat mapping used for?

Rowley: A heat map tries to give you an indication of where liquidity is right now. A statistical engine will pull in lots of inputs and build a heat map of where estimated liquidity is. Then your algos read that and decide where to place their orders.

WS&T: Are your European clients interested in U.S. dark pools?

Rowley: Yes, we get asked about that all the time. A client may have a small basket of orders and they know they're very illiquid stocks. They'll ask us to do something with the dark pools here because they're concerned about their execution quality. We'll put the orders on our dark pool aggregation algorithm, Crossfire. Generally they're very pleased with the results. When executions happen and they don't see the bid and ask move, that's quite powerful for them. One client has called our algo "a beast," which is a good term because it means they're happy with the liquidity.

WS&T: What software tools do you use besides BlueBox?

Rowley: Our quants use Matlab, we've built a lot around that. For instance, we built a market simulator on Matlab. We're prototyping new algorithms in Python.