Quantitative trading Starter

Quantitative trading, refers to the use of mathematical models to find a stable excess returns of investment instruments in the financial markets. Quantitative trading has a strong mining information capabilities, is not susceptible to subjective emotional impact, orders timely, accurate, strict risk control, etc., to obtain steady income. While its investment compared to traditional subjective, difficult to get started is relatively large, the threshold is higher. Getting quantitative trading, mainly requires knowledge of the following aspects.



1. mathematical / statistical knowledge

since it comes with a mathematical model, that knowledge of mathematics and statistics is essential. As the domestic financial market is not yet complete, some of the derivatives trading is limited, compared to foreign markets, to use mathematical / statistical knowledge should be less. For non-engineering background investors need to supplement the knowledge base of the discipline of mathematics, linear algebra, probability, statistics, optimization theory and so on, such as those found in college textbooks. For some new trading strategies using machine learning, data mining also need to know some knowledge. But since it is started, this part of the course, not necessary.

In addition, Applied Econometrics is particularly extensive. Often face a lot of the time series, panel data for policy research. Although more attention to the policy results in practice, as long as money strategy is a good strategy, but with the support of rigorous measurement theory, regression results more accurate and better able to characterize the relationship behind the data, it is often easier to obtain expected results similar. Wherein the cross-sectional time series regression, regression panel logic are very different assumptions, and are widely used in characterization and prediction gains financial assets, fluctuations. Recommended books econometrics Wooldridge's "Introduction to Econometrics: A Modern Approach"; time series recommended "Introduction to financial econometrics" Brooks.



2. programming capability

Since quantitative strategies to deal with large-scale data, and the use of complex mathematical algorithms, so the need to use the procedure to complete this process. Most object-oriented programming languages, such as Python, Java, R and so can do the job. I recommend Python here, in the industry more mainstream, mainly characterized by a large number of packages, including third-party developers, such as Numpy processing of data, Pandas, and financial packages Talib, and compatibility with various platforms and other languages well. Pandas are among the leading US hedge fund AQR development of data processing package, ideal for financial data. Python learning through "the use of Python for data analysis" and other books to learn, you can quick start by some online tutorials. In actual application, each kit should refer more to the API documentation.

Back measuring program and the initialization data includes introducing accounts, each transaction time point selection condition, logic transfer positions, and back to the measurement result of the calculation, the net curve plotted like. Some quantification package backtesting platform environments, simplifies this process, to facilitate the strategy for testing.



3. Financial Basics

quantitative trading, financial market behavior fundamentally. Although there are requirements for the position of mathematics, programming knowledge, but out of its financial nature, we can not design a good strategy. Quantify investors need to understand the nature of various financial assets, as well as factors affecting its price. For equities, the company's fundamentals and financial situation, their situation in which the industry can reflect to some extent in its stock price, investors should have a basic understanding of this. This part can refer to Bodie, Kane, Marcus "Investment", as well as financial accounting, reporting, books. In addition, the Chinese market has been manipulated factors are more significant in firm operations, the quantization investors in making investment decisions rely on quantitative strategies at the same time, the general will add some subjective judgments, in a more timely capture market trends, higher benefits. Therefore, macroeconomic policies affect the situation on financial markets, but also investors can not ignore the problem. Take a look at financial news every day, a long time can develop financial intuition.



4. Policy research capacity

That is the integrated use of the above, the idea of investment procedures, and develop the ability to become strategy of investment value. At the start, you should refer to the existing mature strategy, perfect copy. Logical policy can be summarized in a few words itself may be, but the details in the actual execution can not be ignored. As we all know, outstanding performance in back-tested strategies are not necessarily effective in the real dish, but in the back to test the effect is not a good strategy, it is difficult to have a good performance in the heavy firm. Overfitting, survivor bias and future functions are using the wrong novice would often avoid these errors, in order to make a better backtesting results close to the real situation. Meanwhile, after getting test results back, how that income attribution analysis, equity research positions, risk exposure, and to optimize the parameters, but also to quantify the problem investors need to be addressed.

Some of the classic investment strategies including multi-factor strategy (Fama-French three-factor model), select when technical indicators (MACD, Bollinger Bands, etc.), reversal of momentum strategies, event-driven strategies, statistical arbitrage strategies. Many foreign policy stems from academic, high-quality academic journals including the Journal of Finance, Journal of Financial Economics, and so on. At the same time there are a number of books teaching system, including Barra Handbook (multi-factor Bible), Quantitative Equity Portfolio Management (mainly on portfolio management), Quantitative Trading Strategies (mainly about how to construct quantitative strategies).



5. In practice learning

strategies backtesting after all, is back-tested. Strategies based on past market design, will be able to have a good performance in the past period of time. But the same history does not necessarily repeat itself, with changes in market trends and microstructure strategy may not develop in the direction expected in the coming time. There is also a real problem intraday report information published delay, the transaction friction, single impact on the market price. It is a trading strategy, after a rigorous and comprehensive backtesting test, to test its real effect on the firm.

In early trading contacts quantify, understand mathematical programming model to build the details of the deal are not open around the problem. Today various techniques are relatively mature, available for everyone to use, a successful investor unique place some idea of ​​its design strategy, and in the grasp of the market. Design trading strategies should be based on financial intuition behind, I have been convinced of the idea. I hope investors can find their own unique perspective in the field of investment quantify, to become the next Simmons!