NEHAL PATEL

1
ProActive for Algorithmic Trading

• NEHAL PATEL
• GMO
• OCTOBER 21, 2008

2
Algorithmic Trading at GMO

• What we do Introduction to our business
• High-level grid/distributed computing goals
• ProActive

3
What we do

• Our Business Invest money for clients
• Methodology Algorithmic Trading
• Computers
• Math
• (some ) Finance
• Our Team 12 members based in Boston and Minsk
• Educational background Math, Physics, Computer
  Science
• Work Experience Finance, Academics, Biotech,
  Internet startups

4
Current Environment

• Trades Currencies and futures on 24 hour,
  continuous basis
• 4000 trades a day, 10 billion, 9 banks
• Two data centers, 200 nodes, dozens processes

5
Motivation

• Basic Problem How can you make money using
  computers for investing?
• Basic Wisdom
• Trade frequently, but smartly /Law of Large
  Numbers
• Ideas last for months not years
• Agility, Creativity
• Find the right edge
• Math and CS are equally important
• Embrace complexity, dont sacrifice robustness

6
Methodology

• Mathematics
• Machine Learning/Data mining (Like Biotech, Web)
• Dynamic Programming (Like Airline Scheduling)
• Computer Science
• Java (750,000 lines)
• Emphasis on infrastructure
• Utilize open source
• Data Sets 4-5 Terabytes
• Real-time, Event processing 30K messages/s

7
Algorithmic Trading at GMO

• What we do Introduction to our business
• High-level computing goals
• ProActive

8
Distributed Computing Tasks

• Real-time Stochastic Control
• Complex Event Processing
• Real Time Statistical Analysis
• Data Mining Large Datasets
• Batch Processing
• Historical Simulation

9
Distributed Computing Goals

• Style
• Distributed Objects living in a sea of multicast
  data
• Lightweight
• Agile
• Features
• Scalability
• P2P / Data Push-Pull
• Same Code both simulation and production

10
Algorithmic Trading at GMO

• What we do Introduction to our business
• High-level distributed computing goals
• ProActive

11
Attractive Features

• POJOS
• Separation of Concerns
• High Speed simulation
• Toolset Focus
• Rapid Deployment
• Dynamic byte code

12
ProActive

• Consulting with ActiveEon
• Low Level Architecture
• Topics
• Performance
• More Transparency for POJOs