Friday, March 8, 2013

Book Review - Mean Reversion Trading Systems by Howard Bandy


I am just about finished with Howard Bandy’s new book, “MeanReversion Trading Systems – Practical Methods for Swing Trading”.  While I very rarely review books here on Quantifiable Edges, this one really stands out and deserves some attention.

Howard goes through every step of the systems-building process.  He examines several different oscillators.  He scrutinizes entry & exits techniques.  He discusses risk control.  And on top of it all, he provides code for everything he covers in the book.  It is $50 for the book, which is a ridiculously low price.  There are trading courses that cost many thousands of dollars that don’t provide as much good information as Howard’s “Mean Reversion Trading Systems”.  All of the coding is done in Amibroker, which unfortunately I do not use.  But since he lists it all out, those who use other programs like me can translate it into Tradestation, R, or whatever.  And here is the kicker for anyone that does use Amibroker – Howard has actually set up a web page where book purchasers can download the code at no additional cost.

I commend Howard on his efforts.  If you have an interest in developing your own trading systems, this book is a wonderful resource that I would highly recommend.

5 comments:

Unknown said...

I have been following your blog for a while. But I am now surprised because you commend the work of someone who claims in his book that: (http://www.meanreversiontradingsystems.com/MRTS%20AnalysisWM.pdf)

"My view is that the length of the in-sample period should be as short as is practical. The only way to determine the length of the in-sample period is to run some tests."

This is called data-snooping

He also claims:

"The length of the out-of-sample period is: As long as the model and the market remain in sync and the system remains profitable. There is no general relationship between the length of the out-of-sample period and the length of the in-sample period."

SO we choose the out-of-sample for as long as the model and the market are in sync and the system remains profitable. Very nice work. I wonder why you endorse such stuff. What do you have to gain. Or maybe because I respect your work maybe you overlooked the details. The substance in trading is in the details.

Rob Hanna said...

What a sad world when saying something nice about someone else's work brings emails asking me what I have to gain.

The review got me a nice thank you note from Mr. Bandy, whom I have never met nor spoken to before.

While he views some aspects of testing differently than I, I have no interest in arguing every point he makes in his book. For me, if you can take valuable ideas and information from a book, then it is worthwhile. This one is filled with them.

I stand by my review. I thought the book had lots of great info. It was backed up by actual test results (a rarity), and since he provides all the code, traders can verify the results and easily explore the ideas further on their own.

Those who have read the book are welcome to post comments (positive or negative) below. You all know my opinion.

Rob Hanna

Unknown said...

Instead of feelling sad, maybe you should be happy that someone took the time to point out to you the mistakes in that book which are of fundamental nature, i.e. curve-fititng, optimization, data snooping and all that nonsense that make traders lose money. Don't feel sad. The world is not sad when we go against reality, we should just change course. Thanks.

JerryC said...

I received Howard's book yesterday, and while I haven't finished it yet, I think the 'data snooping' comment is a bit over the top. Howard is constantly cautioning about 'future leaks' and faux optimization techniques. Perhaps mati should actually purchase the book before dissing it at his level.

Unknown said...

I came across this comment and as someone who has all four of Dr Bandy's books, I felt I should chime in on this topic.

Dr Bandy is a strong proponent of good system development practices and his writings clearly warn about the real dangers of curve-fitting. Anyone who has followed his blog or read his book in detail will completely understand the nuance behind his stated views on in-sample/out-of-sample period that one person found fault with.

Dr Bandy has become my favorite author on the topic of quantitative trading approaches.