Interview with Daniel: Treat Data like Gold
-
Daniel won the Q8 and Q9 Quantiacs competitions and got an initial investment of $1,000,000 for each of his winning systems. After that he continued developing systems and he currently gets a monthly fee from Quantiacs. We met Daniel and asked him a few questions about how he got started with Quantiacs.
Q: Can you tell us a little more about yourself and your background?
I’m a mechatronics engineer and hold a MS in engineering. As an engineer, most of my experience is with software, electronics, and hardware. I may have a different background than most users of Quantiacs because I have a lot of coding experience but I spend most of my time trading and providing trading services to financial institutions.
Q: Why did you decide to go into trading?
I was entrepreneurial and had founded a couple engineering companies, but two years ago I decided to change my occupation and move into trading. Trading is related to my previous data analysis work and I was able to find opportunities in the markets.
Q: What markets do you trade?
I trade here in Brazil on the BMF, a commodity exchange, and on the BOVESPA. I trade a small number of instruments, such as two or three currencies, one index, some equities, and options.
Q: How hard was it for you to work with Quantiacs’ data and software?
It wasn’t hard. I am used to work with intraday data up to a tick-by-tick resolution for my own trading. Working with end of day data was easy and the backtests evaluate reasonably fast. If it was with tick-by-tick data I wouldn’t have had the time to enter the competition.
Q: How much time have you dedicated to creating your algorithms?
The Q8 was my third competition, and I certainly spent half a month of full time work. I even took a few days off before the submission deadline so I could spend more time on it.
Q: What was your biggest learning?
I think I added some new market experience to my knowledge. I’m not used to the US and EU futures markets, they are not the markets I usually trade. I gained some knowledge on these markets and that helped me to achieve better results.
Q: What methods do you use for trading systems?
I used machine learning, and I’m not a big fan of technical analysis. I can’t really classify my strategy as trend following or mean reverting. I made my systems very reactive to market changes. Most people tend to focus primarily on data analysis, but I think doing market analysis is equally important.
Q: How do you analyze the markets?
My system only trades a selected set of Futures. Those are the markets I know enough about to trade them. I wouldn’t trade an instrument that I have no knowledge of. My market analysis is primarily based on correlations.
Q: What’s your favorite programming language?
Java is my favorite language, but of course it also has some limitations. For Quantiacs I used Matlab.
Q: How did you find out about Quantiacs?
I was doing research. I was trying to find market data providers for the US markets. I googled for market data and I found Quantiacs right away. I saw the competition and I decided to go for it. It was a challenge to try different markets, and I decided to test my knowledge and methods in the competition.
Q: Can you remember your lowest low and you highest high during the Q8 Competition?
Well, about the lows... Financial markets are very, very hard to predict. They need to be treated as something that’s very hard to beat. This is what discouraged me the most. It’s quite unpredictable. Even working with very sophisticated tools it discourages me, as there is a limit on how much we can do to reach positive results.
The highest high was when I received the news about the first place nomination. I’m even more proud and happy because I think I have now reached a second level consistency. That’s the most important thing for me in the long term. Consistency of the result is most important.
Q: What do you think is the biggest hurdle to get started and be successful on our platform?
I don’t think it’s the technical hurdles. I think it is overcoming discouragement. Financial markets are hard to predict, and not everything you try will work right away.
Q: How did you overcome this hurdle? How do you keep going?
It’s my fundamental belief in math tools, computing capacity and machine learning. I’m very direct in my thoughts. I really believe the financial result is the ultimate goal of the trader. Money tells you if you’re right or wrong.
Q: What other resources would you like to have from Quantiacs?
I work with tick-by-tick data. The more data you have the greater the possibility of success, so it would be nice to have more data. However it is also important to avoid overfitting. The more data you have, the more likely you overfit your system.
Q: How do you avoid overfitting?
One of the things I do is to reduce the data that I use for my predictions. You need to treat data like gold. The more you use it, the more you spend it.
Q: What did your friends say when you won the Q8 competition?
Most of my friends don’t know what a Sharpe Ratio is, but generally they were very happy for me. Even knowledgeable traders and trainers of traders gave me good feedback. Those who know the Sharpe Ratio metric found my SR of 3.6 quite impressive. I would trade that for my own account without hesitation.
Q: What is the most important character trait for a quant developer and why?
Quantiacs’ quants are very good data scientists. It would be interesting to add market knowledge to the data science skills. Try to reach an equilibrium between market and data knowledge. No or little market knowledge is unacceptable for professional traders.
Q: Can you give tips to people who are new to algorithmic trading?
I think new quants need to work on their weaknesses. If someone is good on market knowledge they need to work on data science and vice versa. Make yourself familiar with the programming language, data analysis, or machine learning for instance.
Thanks to Daniel for taking the time for this interview!
