# quantitative trading

This is the third in a multi-part series in which we explore and compare various deep learning tools and techniques for market forecasting using Keras and TensorFlow. In Part 1, we introduced Keras and discussed some of the major obstacles to using deep learning techniques in trading systems, including a warning about attempting to extract meaningful signals from historical market data. If you haven’t read that article, it is highly recommended that you do so before proceeding, as the context it provides is important. Read Part 1 here. Part 2 provides a walk-through of setting up Keras and Tensorflow for R using either the default CPU-based configuration, or the more complex and involved (but well worth it) GPU-based configuration under the Windows environment. Read Part 2 here. Part 3 is an introduction to the model building, training and evaluation process in Keras. We train a simple feed forward network to predict the direction of a foreign exchange market over a time horizon of hour and assess its performance. [thrive_leads id='4507'] . Now that you can train your deep learning models on a GPU, the fun can really start....

This is the first post in our 3-part Back to Basics series. You may be interested in checking out the other posts in this series: Part 2: How to Succeed at Algorithmic Trading Part 3: Backtesting in Algorithmic Trading This is the first in a series of posts in which we will change gears slightly and take a look at some of the fundamentals of algorithmic trading. So far, Robot Wealth has focused on machine learning and quantitative trading research, but I had several conversations recently that motivated me to explore some of the fundamental questions around algorithmic trading. In the next few posts, we will investigate questions such as: What is algorithmic trading? What can algorithmic trading do for me? What are the pre-requisites? What should I think about before getting started? What's all this fuss about curve fitting and robust optimisation? Why should I care? So without further ado, lets dive in! What is Algorithmic Trading? At its most basic level, algorithmic trading is simply the automated buying and selling of financial instruments, like stocks, bonds and futures. It requires a...

This is the first post in a two-part series about the Hurst Exponent. Tom and I worked on this series together and I drew on some of his previously published work as well as other sources like the very useful Quantstart.com. UPDATE 03/01/16: Please note that the Python code below has been updated with a more accurate algorithm for calculating Hurst. Thanks Mike at Quantstart.com. Mean-reverting time series have long been a fruitful playground for quantitative traders. In fact, some of the biggest names in quant trading allegedly made their fortunes exploiting mean reversion of financial time series such as artificially constructed spreads, which are used in pairs trading. Identifying mean reversion is therefore of significant interest to algorithmic traders. This is not as simple as it sounds, in part due to the non-stationary nature of financial data. We both think that Ernie Chan's book “Algorithmic Trading: Winning Strategies and Their Rationale”, is one of the better introductions to mean reversion available in the public domain. In the book, Ernie talks about several tools that can be used when testing if a time series...

This post comes to you from Dr Tom Starke, a good friend of Robot Wealth. Tom is a physicist, quant developer and experienced algo trader with keen interests in machine learning and quantum computing. I am thrilled that Tom is sharing his knowledge and expertise with the Robot Wealth community. Over to you, Tom. Unlike most other businesses, algorithmic trading has the advantage of giving you almost instant feedback on how good you are in your business. For anyone who is numerically inclined this is a very attractive proposition. I have seen articles written about this subject but they have never really addressed a lot of the issues I have come across on my journey. In this post I would like to talk about this a little as an inspiration or perhaps a deterrent for all the people who read this and consider making money that way. Nothing could be more amazing, a system that runs by itself and consistently spits out cash to finance prolonged stays in Bali, South America or with your mom if that’s what you’re after. However,...

In the first post in this series, I explored mean reversion of individual financial time series using techniques such as the Augmented Dickey-Fuller test, the Hurst exponent and the Ornstein-Uhlenbeck equation for a mean reverting stochastic process. I also presented a simple linear mean reversion strategy as a proof of concept. In this post, I’ll explore artificial stationary time series and will present a more practical trading strategy for exploiting mean reversion. Again this work is based on Ernie Chan's Algorithmic Trading, which I highly recommend and have used as inspiration for a great deal of my own research. In presenting my results, I have purposefully shown equity curves from mean reversion strategies that go through periods of stellar performance as well as periods so bad that they would send most traders broke. Rather than cherry pick the good performance, I want to demonstrate what I think is of utmost importance in this type of trading, namely that the nature of mean reversion for any financial time series is constantly changing. At times this dynamism can be accounted for by updating the hedge...

Important preface: This post is in no way intended to showcase a particular trading strategy. It is purely to share and demonstrate the use of the framework I've put together to speed the research and development process for a particular type of trading strategy. Comments and critiques regarding the framework and the methodology used are most welcome. Backtest results presented are for illustrating the methodology and describing the outputs only. That done, on to the interesting stuff My last two posts (Part 1 here and Part 2 here) explored applying the k-means clustering algorithm for unsupervised discovery of candlestick patterns. The results were interesting enough (to me at least) to justify further research in this domain, but nothing presented thus far would be of much use in a standalone trading system. There are many possible directions in which this research could go. Some ideas that could be worth pursuing include: Providing the clustering algorithm with other data, such as trend or volatility information; Extending the search to include two- and three-day patterns; Varying the number of clusters; Searching across markets and asset...

In the last article, I described an application of the k-means clustering algorithm for classifying candlesticks based on the relative position of their open, high, low and close. This was a simple enough exercise, but now I tackle something more challenging: isolating information that is both useful and practical to real trading. I'll initially try two approaches: Investigate whether there are any statistically significant patterns in certain clusters following others Investigate the distribution of next day returns following the appearance of a candle from each cluster The insights gained from this analysis will hopefully inform the next direction of this research. Data preliminaries In the last article, I classified twelve months of daily candles (June 2014 - July 2015) into eight clusters. To simplify the analysis and ensure that enough instances of each cluster are observed, I'll reduce the number of clusters to four and extend the history to cover 2008-2015. I'll exclude my 2015 data for now in case I need a final, unseen test set at some point in the future. Here's a subset of the candles over the entire price history (2008-2014, 2015...

This post builds on work done by jcl over at his blog, The Financial Hacker. He proposes the Cold Blood Index as a means of objectively deciding whether to continue trading a system through a drawdown. I was recently looking for a solution like this and actually settled on a modification of jcl's second example, where an allowance is made for the drawdown to grow with time. The modification I made was to use the confidence intervals for the maximum drawdown calculated by Zorro’s Monte Carlo engine rather than the maximum drawdown of the backtest. The limitation is that the confidence intervals for the maximum drawdown length are unknown – only those for the maximum drawdown depth are known. I used the maximum drawdown length calculated for the backtest and considered where the backtest drawdown depth lay in relation to the confidence intervals calculated via Monte Carlo to get a feel for whether it was a reasonable value. Below is a chart of the minimum profit for a strategy I recently took live plotted out to the end of 2015, created using the method...