Martin Mayer-Krebs
With an academic background in economics and financial engineering, in addition to having lifelong passion for programming, I work as the lead quantitative developer at Isola Investments.
Before working as an algorithmic developer, I worked as a mathematical modeler at a boutique consulting firm catering its services to the WB, UN, and the IDB, amongst others.
In my free time, I like to code small projects ( like this one), lose chess games ( chess.com), and play the drums.
Articles
To organize my thoughts and find holes in my understanding, I like to write things down. It not only is a great way to learn, but it also has a nice byproduct: I end up writing a few articles. From time to time, I even write one that I like. What follows is a list of my favorite ones:
Bulding a Backtesting Framework in Python
In this article, I demonstrate how to build a backtesting framework in Python from the ground up. The process includes creating core components like the Engine, Strategy, Order, and Trade classes, and enhancing the framework with performance metrics. In the second article, I expand the framework by adding order types and output metrics.
Link to Part I / Link to Part II
Walkforward Optimization in Python
In this article, I explain walk-forward optimization in Python, detailing how to implement and optimize a trading strategy using Backtesting.py. I break down the process into steps, including implementing the strategy, optimizing parameters, and performing walk-forward optimization. This method ensures robust and adaptive trading strategies by iteratively adjusting to new data. I would have liked to make it compatible with my backtesting framework, but I wrote this article before.
Overfitting in Backtesting & Trading
Here I explore the concept of overfitting, explaining how it leads to misleading performance results by fitting a model too closely to historical data. I provide examples, discuss related issues like p-hacking, and offer methods to detect and avoid overfitting, such as out-of-sample testing and ensuring parameter stability.
Most Important Papers for Quantitative Traders
The title is very misleading, as I actually list my personal favorite papers. Covering a range of topics from optimal execution algorithms to behavioral finance, these papers provide foundational knowledge for anyone interested in the field.
Is Backtesting Accurate?
Since I wrote extensively on backtesting, I also went ahead and wrote a short article (disclaimer) on the dangers and pitfalls of backtesting. It might seem too simplistic, but it addresses 99% of the mistakes I’ve come across.
Personal Projects
Over the years I’ve coded plenty of small tools and projects, but I won’t list them here because they’re mostly unrelated to quantitative finance
QMRExchange
The qmrExchange project is an open-source financial markets exchange simulator that realistically mimics all the main components of modern trading venues. It allows us to test and quantify the behavior of different agents in a laboratory and isolated environment without the high noise-to-signal ratio that is otherwise unavoidable in live settings. By creating a completely functioning trading venue whose access is only granted to a finite and known number of agents or trading algorithms, qmrExchange enables analyzing causation and quantifying the impact of each agent in a way that is otherwise unfeasible.
I’m probably biased, but I think the project is really interesting. At the very least, I haven’t seen any other project that does the same.