Title: Designing Stock Market Trading Systems (with and without soft computing)
Authors: Dr Bruce Vanstone & Tobias Hahn
I've been eagerly awaiting this book after ordering it from fishpond.com.au ($80). It arrived yesterday, and i finished reading it in about 4 hours. That's $20/hr just in case you didn't pick up on it. There's your first clue as to my impression.
So. Very. Disappointed.
To the authors, i apologise for what follows. But, you have one seriously annoyed customer here.
I was surprised at how small it was when i first saw it (~240pgs). I was sure it was going to be far more substantial given the price tag (note the RRP is $160!). In fact, had i picked this up in a book store, i definitely would not have bought it simply based on this. I'll be more careful about this when buying online in the future.
After some sage advice from a good friend i tried not to let these first impressions sour my opinion.
Unfortunately the content wasn't any better.
I truly wanted to like this book. It offered much, and i was drawn to the following:
- It ticked all the criteria i look for in contents (risk, position sizing etc).
- It was very recently published - 2010
- It promised to show you how to test published research
- It discussed applications of soft computing
- The authors are both Australian (although i didn't realise this at the time of purchase)
I found the best chapter was that describing soft computing. There was a (brief) discussion on different algorithms, and a fairly useful summary of some research conducted in the area. But even this i found myself skim reading after a while.
I was surprised that what i perceived to be the "centre piece" chapter on Artificial Nueral Networks (ANNs) was lacking. The authors have done a great deal of research in this area, so i was expecting them to really deliver here. But after reading this I really had no better idea how ANNs work. In fact, i found myself reaching for my copy of "New Trading Systems and Methods" by Kaufman which did a far better job of explaining how they work.
This can be extended to the rest of the book. Sure you can't describe everything, but i really think i should come away with some further understanding.
If this was all that was wrong it could be forgiven. But my biggest gripe by far was the poor quality of the presentation. Examples:
Plots with axes not labelled, and no discussion in the text. From my report writing experience this is a big no-no.
Some tables were very difficult to read. No row lines, and content didn't line up on the rows.
I believe too much space was used to describe the process used to test things in specific software (Neuro-lab by Fidelity Investment) which i'm suspect few would be interested in using.
Equations presented without any description of how to interpret the results. This was the case for many of them. Example:
f = ((b + 1)*p - 1)/b
p = probability of winning
b = ratio of amount one to lost"
That's it…. What about 'f' !! I found no description of what 'f' represents (i.e. proportion of capital to bet). Luckily for me i've come across this before, but with other concepts i wasn't so lucky. For the ANOVA (analysis of variance) equations i couldn't even interpret all the functions due to typos, and missed spaces.
All this is not to say i got nothing out of reading the book. It did spark a few ideas, and i'm keen to investigate ANNs further. But if i buy a book like this i expect to come away having a better understanding of the topic than when i started. Especially for a fairly hefty $80.
It was the really very poor quality in how the content was presented that has led me to write this poor recommendation. If a report of this quality was given to me at work i would be inclined to tell the author to bring it back after they'd finished.
I recommend saving your money and instead downloading their research papers from the web: trading.it.bond.edu.au - you'll get much the same (if not more) information out of them.
I really don't want to bad mouth people. I even specifically waited a day before writing this so i would calm down. I'm sure the authors are highly intelligent, and successful at their research and trading ventures. But i found the quality and lack of explanations simply inexcusable for a book of this price.