GitHub Financial Project

is launched!

 

 

 

 

 

 

 

Features:

 

- Free High-Quality Financial Data.

 

- Over 200 GB Financial Cloud Data Delta Lake.


- Easy for Everyone.


- Providing historical S&P 500 constituents high-frequency data (No survivorship bias), and alternative data. Data is

  back to 2004.


- 100x times faster than traditional database using Koalas, Apache Spark.


- Specially designed for feature selection and Machine Learning.

 

 

 

 

About Us

 

Matroid Evolved is a financial research firm dedicated to fighting against p-hacking and financial fraud.

We manage clients' assets by using the most advanced tools.

 

Approach

We use statistical and mathematical inference to invest in financial markets, with our decisions driven by data and economic theory.

 

 

 

Technologies

We use different simulations to prevent overfitting and false discovery.

 

 

 

 

 

Opportunities

Our people come from a board range of backgrounds and disciplines.

 

 

 

 

 

Why us?

 

 

 

Financial data sets exhibit a lower signal-to-noise ratio. One of the most pervasive mistakes in financial research is to take some data, run it through a Machine Learning algorithm, backtest the predictions, and repeat the sequence until a nice-looking backtest shows up. Academic journals are filled with such pseudo-discoveries, and even large funds constantly fall into this trap. The fact that we are repeating a test over and over on the same data (We call it p-hacking) will likely lead to false discovery. This methodological error is considered by some statisticians as a scientific fraud, and the American Statistical Association warns against it in its ethical guidelines.

 

 

 

Future Performance of Most Financial Products in Reality

Source: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3544431

 

 

 

One example is the increasing prevalence of backtest overfitting, due in part to the ease of generating large numbers of model variations using modern computer technology. Very few peer-reviewed research papers or commercial products disclose the number of computer trials used in the development, so it follows that many published and marketed strategies are statistically bogus. Indeed, such statistical errors are the primary reason that investment funds and strategies, designed for optimal performance based on historical market data, often fail when actually fielded. Mathematicians Against Fraudulent Financial and Investment Advice (MAFFIA) also warns it and raises public awareness of the problem.

 

Unlike most financial firms, we persist in using the scientific gold standard to do financial research in order to prevent overfitting and bias to occur so that we can deliver good results for our clients. Our experiments are transparent and repeatable, so you should be able to obtain the same result. Even, our trading strategies are much more transparent than most firms. 

 

The following video is about how most financial firms do research and trading.

 

 

 

 

Collaborators