I've just finished reading the somewhat grandiosely named "Statistically Sound Machine Learning for Algorithmic Trading of Financial Instruments: Developing Predictive-Model-Based Trading Systems Using TSSB" by David Aronson and Timothy Masters. This is the companion book to their free software, TSSB, for developing machine learning models for trading. The book cost about $120 from Amazon, but it was well worth the investment. There are some fantastic ideas contained within and the book is essentially a user manual for TSSB.

As a future feature, I would like to suggest a bridge between Zorro and TSSB. I think that the two could be an extremely powerful team. For example, Zorro could access TSSB's machine learning algorithms in a similar manner to the R bridge.

Even more importantly in my opinion, TSSB is able to create and manage large databases of price derivatives and other data, which facilitates the systematic mining of trading ideas.

Coupling the power, simplicity and robustness of Zorro with TSSB's feature management capabilities would be extremely useful.

Has anyone else looked at TSSB or read Aronson's latest book? I would love to hear your thoughts. I am currently working on implementing Aronson's idea for a machine learning algorithm that acts as a trade filter rather than a price movement forecaster. I'm seeing significant out of sample performance improvement with this method and would really like to discuss my approach with others who have attempted something similar. I'm using R instead of TSSB, but am basically using Aronson's approach.