The Algorithmic Trading Podcast

ATP02 - Stuart Bevan, Credit Suisse

11.19.07

Stuart Bevan
Stuart Bevan, Credit Suisse

Welcome to episode two of The Algorithmic Trading Podcast, brought to you by Voices in Business and sponsored by Sybase.

Today, we feature Stuart Bevan, Head of Alternative Execution Products Technology at Credit Suisse, in conversation with Sinan Baskan, Senior Product Manager at Sybase and Greg Grimer of Voices in Business.

Detailed show notes for this episode:

00:13 - Introduction
00:56 - Start of interview
01:02 - The impact of MiFID & RegNMS on data distribution
03:13 - Breakthrough technologies
04:12 - Competition from smaller niche players
05:18 - Growth in complexity
07:08 - Growth in volume of data & events
08:02 - Shared services & fragmentation of data
09:24 - Development issues & toolkits solutions
12:26 - US vs Europe vs Asia
13:26 - Complex event processing
14:40 - Re-usability & lifecycle of custom developments
16:45 - What would be the ideal technology?
18:28 - Re-usability of data & anonymity
19:39 - Security issues & performance vs functionality
22:30 - Developer skills & program management
25:08 - End of interview & wrap-up

Click here for a full transcript of this episode.

We welcome all feedback, so please leave a comment here on the website, call us to leave an audio comment at +44 (0) 20 7193 1295, or send a message to algo@voicesinbusiness.com.

Listen Now:

To listen to the podcast right now, click on the player icon below.

 
icon for podpress  ATP002 - Stuart Bevan, Credit Suisse [25:55m]: Play Now | Play in Popup | Download

3 comments so far

Nowadays, we have a lot of professionals that use their skills to build profitable automated trading systems. Let’s take metatrader platform for instance, which is free and easy to code for. I believe that very soon we would have much more sophisticated libraries available for programmers (even now metatrader code can be easily integrated with fuzzy logic, NN, non-linear dynamic tools and anything available from Mathlab or other tools), so more market-sensitive trading code will emerge. The next step is computational speed (maybe we’ll have tools for distributed computations as they have now in 3D MAX package, but for trading agents
optimization) - the cost will become even cheaper. Even now a typical person could rent, let’s say, 20 machines in an internet cafe for distributed computations. Perhaps for $500 or less.

So the real intriguing question is - HOW FAR CAN WE GO LIKE THIS? If even now one (common person, not a professional trader) may buy a profitable (more or less…) mql script for like 30->2000$ and make money, then if these systems become cheaper and better in quality - all would be able to make money? But how could that work???

what I’ve thought of, perhaphaps there’s a way to model this situation using systems analysis software. Suppose we model fruits and wegetables bazare, there natual cycles and people that sell and buy (adding little white noize or smth like this), and then we introduce a genetically bread agents that analyze the prices and trade futures on these market and make these agents compete (only the agents that perform well according to some given benchmark) - and see if its really possible to have all these systems when everyone makes money and happy…

or perhaps these simualtions exist already…



Leave a comment
Line and paragraph breaks automatic, e-mail address never displayed, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>

(required)

(required but not displayed)