Michael's blog

Backtesting is the most important stage of testing and verifying the result of your trading strategy. You can either build your own backtesting script or use a python package such as backtrader to simulate your portfolio return with downloaded stock prices CSV files. Either way would require you to download the stock price in minute or day bar beforehand. Also, you need to deal with the stock market events such as stock splitting, ticker name changing, or delisting. So effectively using an existing online tool like QuantConnect could save you a lot of time dealing with edge cases yourself.

I’ve been spending my time learning how to work with QuantConnect platform and its features in the past month. So in this post, I’m going to introduce the web-based backtest platform by finding out the common things between Quantopian and QuantConnect. Then I’ll talk about the good and bad I found in trying out these two platforms. Hopefully, the experience that I shared in this post can better help people who are looking for a solution of validating their trading ideas.

In the post 【Machine Learning】 Part II - How to build a machine learning boilerplate?, we have successfully built our machine learning boilerplate. By having this template, we can develop an advanced machine learning trading strategy upon. However, even with the strategy result that looks profoundly profitable, we still won’t be able to know how much money we can make by looking at the accuracy rate of our machine learning trading algorithm.

In order to better understand whether the results from the output of our model are really concerning our portfolio return, I put together a rather simple strategy and run several backtests with different parameters. In the end, we’re going to answer several frequently asked questions in order to decrypt the myths of machine learning trading algorithms.

To follow up on what we have learned in the last post, here we’re going to quickly demonstrate how to build your own machine learning boilerplate. We’re going to cover the steps that we introduced in the post:

1. Data curating
2. Feature discover/analyze
3. Train the model
4. Predict the expected variable
5. Form the strategy and run backtest
   Read more »

Machine learning has been a black box for most people. We have been told how powerful it is to address our questions with the terms such as CNN, LSTM, training data, modeling, …. Also, machine learning has been practicing in so many fields such as medical, finance, macroeconomic, marketing, …etc to address the most difficult and complex questions.

Using machine learning as a tool to beat the market would be our goal here. Therefore, I’ve put together a list of questions that I used to have and will address them one by one in this post, hoping this would help people who have the same questions as I did. The fundamental ideas and theories behind machine learning are all about math, which I’m not going to over in this article as it would take ages to explain them in a better way than all the other existing articles that you can find on the internet.

The questions are:

• What does Machine learning do in the finance world?
• How does machine learning work?
• What are the FAMQ (Frequently Asked Myself Questions)?
• Postlogue: What is our ultimate goal of developing a quantitative investment strategy?
  Read more »

It’s almost been a year I started working as a self-employed quant trader. I’ve been learning, researching, and coding most of the work time, then I invest to the rest of my time to taking care of families and tried to have quality time with them during this COVID year. Even though I broke my knee ligament while playing basketball and rest on the bed for almost two months, I’m still coding and refactoring my automatic trading framework in order to ready to dry run my next ready-to-go strategy.

However, I think I owe myself a transcript of the year to make sure I haven’t slacked off in this past year. Also, I need to set up a few goals in the new year.

This is the finale of the【Factor analysis】 series. After all the things we have done: downloading the data, generating factor data, and conduct factor scoring, now we’re in the last part to evaluate whether our deeds are effective enough to profit.

Before analyzing anything, you need to have something to be analyzed. It’s unlikely that we’re going to buy all the good quality stocks, nor we’re able to define a fine line that above which score will make us money. So we have to build a strategy that we deem it’s profitable and then analyze the strategy against the benchmark portfolio, which is S&P 500 index.

Previous reading

• 【Factor analysis】 Vol. 1. Introduction the idea of factor analysis
• 【Factor analysis】 Vol. 2. Get ready - preparing and cleaning data
• 【Factor analysis】 Vol. 3. Construct your own factor

After we have collected all the fundamental data and have produced the factor data we need, how we can utilize them to compete with the market and to beat the benchmark return?

One of the common methods would be factor score. In this post, we’re going to describe and implement this factor score method to see how we can rate the stocks that have more potential to rise than the others in the future.

Previous reading

• 【Factor analysis】 Vol. 1. Introduction the idea of factor analysis
• 【Factor analysis】 Vol. 2. Get ready - preparing and cleaning data

After we got the fundamental data we needed from any data source you preferred or paid as in the previous post, we’re now going to generate factors to be used in the factor analysis.

We’ve talked about the definition of survivor bias and the impact if we didn’t notice and resolve it in 【How 2】 Vol. 3. How to produce a quality tradable securities for backtesting. One of the most critical topics that we need to address is to get the accurate list of components of your target portfolio over time when developing your investing strategy. Therefore, I’m gonna use S&P 500 Index as the example to go through this post, and we’re gonna use the result to the later series in 【Factor analysis】 Vol. 1. Introduction the idea of factor analysis.

In the last post Vol. 1. Introduction the idea of factor analysis, we have introduced the concept of factor analysis. There are various ways to implement the factor analysis against the problem you’re trying to resolve or explain. However, any of them would require proper data to be provisioned.

Just like the cover image that I used in this post implies. All the top-tiered chefs know how to make good dishes as their skills are polished and proficient. In the end, the quality of the ingredients matter. That’s why it’s essential to acquire the right data in the right way, and then you’ll have the solid foundation to proceed with the analysis.