Blitzweekend, MandelBrot (3)
The MandelBrot project was 2nd when counting votes from attendees.
Can you present yourself? What is your background?
Ismail Negm: Ismail is a licensed options broker with nearly 5 years of experience in the industry, and is an experienced trading systems developer with a keen interest in risk management strategies. He is also completing the following designations: Derivatives Markets Specialist, Chartered Financial Analyst, Professional Risk Manager and Chartered Market Technician.
Julian Squires: Julian has been working professionally as a programmer for over a decade, and has been involved in the open source community for even longer. His professional interests include pure mathematics and optimizing compiler design.
Can you describe what is the MandelBrot project? How did you get the idea? What problem does it aim to solve?
The idea for the Mandelbrot project (working title) came through our dissatisfaction with standard methods of testing trading strategies, and specifically with the optimization phase of testing. It is standard to test a strategy using historical data for the security that is intended to be traded and to proceed with a walk-forward analysis using the optimized results, again on historical data. Using this standard method yields highly curve-fitted results which are not representative of the future performance of the system in live trading. Ideally, a trader would have access to an infinite amount of representative data on which the strategy can be tested, and that is what our project aims to produce.
What were your objectives regarding Blitzweekend? What did you expect from the event?
Our expectations were very slim: a suitable environment in which to work. Our objective was to build a framework for realistic synthetic data generation. To test our data, we also had to develop a simulation environment that could test various trading strategies on both historical data and our generated synthetic data. We achieved these objective using technologies that yielded performance and development time wins.
Can you talk us about your experience of Blitzweekend? any challenges? How did you overcome problems?
The Blitzweekend experience was wonderful in many respects. We achieved our goal in less time than was allotted to us. We also found that we had access to useful experts and that we were well fed and taken care of. The main challenge was breaking away from all the interesting people we met over the weekend to actually get more work done. We were not successful in overcoming that “problem”.
What is now your plans for MandelBrot?
We are continuing to develop an efficient, scalable commercial product that runs on a fractal-based data generation model.
They got my vote on the product front. I do agree with the choice of Blitz Director as a winner, because both the product and presentaton were amazing, but on the product front, I think Mandelbrot is 1) pure genius and 2) an gargantuan task for this team to have tackled and wrapped their brains around given the 48 hours time constraint.
Oh, and they should have gotten browny points for their choice of platform: http://www.jsoftware.com/
i voted for mandelbrot — it’s an AAA project - although i am not even sure i can grasp what they developed technologically during the weekend
blitzdirectory was also interesting, but on a pure business aspect.
Leave a Reply